Innovation barriers across firm types and countries
Werner Hölzl and Jürgen Janger1
Paper presented at the DIME Final Conference, 6-8 April 2011, Maastricht
Preliminary draft
Abstract
This paper studies the differences in perception of innovation barriers of innovative and non-innovative firms for 18 EU
countries using Community Innovation Surveys for the years 2002-2004 and 2004-2006. The results confirm that the
perception of entry barriers is very different between non-innovators that are hindered in taking up innovation
activities and non-innovators that are not interested in innovation activities. While the share of innovators is
decreasing with the distance to the technological frontier, the share of barrier-related innovators is increasing. The
econometric results show that firm size is generally negatively associated with the perception of higher barriers. Firm
expansion is generally associated with a higher perception of obstacles to innovation. International activity generally
increases the perception of barriers to innovation. With regard to differences across country groups there is a clear
indication that barriers related to the availability of skilled labour, innovation partners and knowledge are more
important as barrier to innovation for firms located in countries close to the frontier, while the opposite is true
regarding the availability of external finance. The evidence regarding barrier-related non-innovators suggests that
there is a negative relationship between the propensity to innovate and the perception of innovation barriers. The
patterns of perception of barriers of barrier-related non-innovators are in line with the overall perception of
innovation barriers in the country.
1
Address: Austrian Economic Research Institute (WIFO), Arsenal Objekt 20, A-1030 Vienna, Austria.
E-mail: [email protected], [email protected].
– 2 –
1. Introduction
Innovation is increasingly seen as a key source for sustaining economic growth and welfare.
Thus many governments enacted policies that are geared towards providing incentives for
firms to engage in innovation activity. A more traditional approach provides subsidies for
innovation-active firms in order to reduce the private costs of innovation and to match the
private returns and the social returns to innovation activities. In the past decades SMEs
increasingly entered the horizon of innovation policy makers. Much innovation policy
measures are directed towards alleviating barriers to innovation especially for small firms.
Increasing the pool of active innovators is today an explicit goal of innovation policy
strategies in many countries. This moved innovation barriers into the center stage of
innovation policy.
While innovation studies have analyzed the determinants of innovative activities at the firm
level, especially emphasizing the technological and organizational capabilities and
associated firm strategies required to become successful innovators. However, the research
did not focus in a systematic way on the determinants that deter innovation activities in firms.
From an innovation policy perspective that aims at increasing the number of potential
innovators, it is important to know which barriers are especially relevant for potential
innovators, in order to put policies in place that foster innovation-based competition and
relax market and system failures to innovation. This paper aims at providing a first step
towards the identification of deterring barriers to innovation in a cross-country context. We
provide evidence on revealed barriers to innovation for innovative firms and deterring barriers
to innovation to non-innovative firms following D’Este at al. 2008. The distinction between
obstacles to innovation that are revealed by innovative firms and deterring barriers to
innovation by non-innovators is necessary to identify hampering factors from entry barriers to
innovation.
This research uses two waves of the Community Innovation Survey (in the years 2004 and
2006) in order to study differences between country groups with regard to revealed and
deterring barriers to innovation. By using innovation survey data for 18 European countries we
are able to study the answers in a different set of countries. These countries are quite different
in terms of their economic and technological development. The 'distance to the frontier'
approach acknowledges the specific role of 'appropriate institutions' (Aghion and Howitt,
2006) at different stages of development. The emphasis is on the argument that the
effectiveness of economic policies is conditional to a country's distance to the world
technological frontier. Using a stylised model, Aghion et al. (2006) show that high-skilled
personnel and technology-intensive firms are more important to economic growth in
countries that are close to the technological frontier than for countries further from the
frontier. Our primary motivation for the research is the expectation that technological and
institutional environment affect the perception of barriers and obstacles to innovation.
– 3 –
The paper is organized as follows: The next section provides a background discussion for the
research. Section 3 presents the data. Section 4 presents method and results. Section 6
concludes the paper.
2. Innovation barriers
Innovation activities are a important element of firm strategies, its performance and its
survival. Both research and policy making have emphasized the role of innovation for
fostering competiveness and sustainable development. Nevertheless, available evidence
indicates that even in the most advanced countries many firms are not involved in
innovation. The evidence in Table 1 shows that even in the most advanced countries of the
EU the majority of firms does not engage in innovation activities.2
This raises the question whether barriers to innovation hinder potential innovators to take up
innovation activities and if so which obstacles are relevant. Potential failures may be related
to external barriers to innovation such as the lack of availability of finance for high risk and
uncertain innovation activities, the lack of technological knowledge and market
opportunities for innovation, a lack of connectivity in the innovation system that does not
provide innovation partners or weaknesses in the supply of an adequate skill-base from
secondary and tertiary education. In addition innovation barriers can be internal to a firm.
Table 1: Innovators across country groups
Full sample
Innovators
35%
Country group 1 Country group 2 Country group 3 Country group 4
45%
34%
37%
20%
R&D innovators
16%
29%
16%
17%
3%
Non-technological
innovators
19%
16%
18%
20%
17%
65%
55%
66%
63%
80%
Non-innovators
Source: CIS 4 and CIS 2006 data accessed at the Eurostat safe centre. WIFO calculations. The numbers are simple
averages over CIS-4 and CIS-2006 averages. See section 3.2 for details on the country groups (group 1: member
states close to technological frontier, group 2: advanced catching up member states, group 3: Southern European
member states with low- to medium tech industry structure and high GDP, group 4 : trailing catching up member
states). Country group 1: Belgium, Denmark, Germany, Finland, France, Iceland, Luxemburg, Norway, Sweden;
Country group 2: Czech Republic, Estonia, Hungary, Slovenia, Slovak Republic, Ireland; Country group 3: Spain, Italy,
Portugal, Greece; Country group 4: Bulgaria, Lithuania, Latvia, Romania, Cyprus, Malta.
As this evidence is evidence is based on Community Innovation Survey Data it is likely to overestimate the share of
innovators. Microenterprises are undersampled in the CIS (cite). Microenterprises have a lower propensity to innovate
and make up the largest share of firms in all European economies (e.g. Hölzl and Reinstaller, xxxx)
2
– 4 –
2.1 External and internal barriers to innovation
It is important to note that our contribution focuses only on barriers to innovation external to
the firm. Many important barriers to innovation are found within the enterprise. Adopting new
technologies, introducing new products and organizational structures creates resistance
within the firm. While internal and external changes often stimulate innovative exploration,
internal resistance to change often prevents it. Within larger and established firm many
different barriers to innovation that affect negatively a firm's ability to create radical
innovations can be identified (Assink 2006):
1. Adoption barriers that are related to dominant designs, path dependency and
successful products limit the ability to search for new disruptive innovations. Such
adoption barriers are often increased by excessive bureaucracy in large enterprises
leading to a status-quo bias status-quo bias where deviations from the standard are
perceived as negative.
2. Mindset barriers that are related to the inability to unlearn the old logic of how
products and markets work. This may also associated with the lack of distinctive
competencies to detect and to exploit opportunities arising from external changes.
3. Risk barriers are associated with an excessive reliance on routines and experience
and an unwillingness to cannibalise the own product markets. Disruptive innovations
often threaten the existing products of established firms.
4. Nascent barriers that are associated with management capabilities to foster thinking
out of the box and the management of the innovation process.
These barriers are internal to the firm and are closely related to the specific management
and organization of a firm. These barriers do not necessarily imply that radical innovation
cannot take place within the firm but they indicate that existing organizations try to resist to
changes, which need not be a bad thing. Not every innovation project is worth being
executed. Innovation barriers can thus also be considered as organizational screening
devices to filter worthy innovation projects from unworthy ones. Tang and Yeo (2003) argue
that such internal barriers may even lead to an improvement of the innovation performance
of enterprises. This shows that innovation barriers do not indicate barriers to innovation but
that they need to be considered as factors that affect the innovation process within
enterprises, deterring, delaying or changing innovative ideas and innovation projects (Mirow,
Hölzle and Gemünden 2007).
External barriers to innovation in contrast are related to the institutional and the market
context and are thus closely associated to market, government and system failures. While
internal barriers to innovation are primarily an issue of management, organization and firm
competences, external barriers emerge when the firm interacts with other firms, agents or
institutions in the economic and innovation system. Issues such as standardization, regulation,
financing of innovation, availability of skilled labour and technology transfer decrease the
incidence of external barriers to innovation to firms with high-potential innovation projects
and form the basis for policy measures to foster the innovation potential of an economy.
– 5 –
While the evidence on internal barriers to innovation is of interest to policy makers as it helps
to understand how firms actually innovate and how external barriers affect the innovation
potential of firms, only evidence on external barriers to innovation provides a basis for policy
intervention. The basic rationale for this view – that external barriers are relevant as policy
rationale but not internal barriers to innovation - is that unexploited innovation opportunities
by large firms are often taken up by innovative entrepreneurs and start-ups.3
2.2 The perception of obstacles: Deterring vs. revealed barriers to innovation
At the firm level, it is well known that the characteristics of the firms affect the perception of
barriers to innovation. Arundel (1997), Mohnen and Rosa (2000), Baldwin and Lin (2002), Galia
and Legros (2004) and Iammarino et al. (2007) show that innovative firms attach higher
importance to the hampering factors to innovation than non- innovators. In addition within
the group of innovating firms the obstacles were considered more relevant firms having high
innovation and R&D intensities. The positive link between innovation intensity and the
propensity to evaluate as important barriers to innovation is less surprising when the original
question of the CIS is considered, that emphasizes hampering factors not barriers. Therefore in
the empirical literature the answers are generally considered as firms' assessment of the
obstacles and as a measure of their ability to overcome them. Baldwin and Lin (2002) and
Galia and Legros (2004) provide two possible complementary interpretations:
1. Performing innovation activities increases the awareness of the difficulties
encountered, without preventing firms' to pursue innovation projects.
2. The formulation of the CIS question on obstacles leads firms to assess the problems
they faced and have overcome in performing innovation activities.
Important from a policy perspective is that not much is known about the barriers to
innovation and the extent to which barriers actually deter the take-up of innovation by noninnovative firms. The existing literature on barriers to innovation has concentrated on the
perception of barriers among innovative firms (e.g. Mohnen and Rosa 1999) or treated noninnovative firms as an undifferentiated group (Hölzl and Friesenbichler 2009, Iammarino et al.
2009) but does not address the core of the policy question which barriers to innovation are
crucial in inhibiting the take-up of innovation activities by non-innovators.
It is thus fair to say that the extent to which barriers actually deter the take-up of innovation
by non-innovative firms is largely unknown, with the important exception of D’Este et al.
(2008), Savignac (2008) and Mohnen et al. (2008). D’Este et al. (2008) are able to show that
non-innovators that have not much interest in performing innovation activities rank obstacles
to innovation very low. However, those non-innovative firms that aspire to be innovative
experience barriers in the same way as innovative firms. Thus they are able to distinguish
3 This shows that there is a close relationship between entrepreneurship policy and innovation policy. In fact, the
promotion of high-technology entrepreneurship and early stage venture capital is today an important element of
innovation policy.
– 6 –
between revealed barriers to innovation and deterring barriers. The first are barriers that
obstruct firms’ achievement in innovation activates, the second type of barriers prevents firms
from engaging in innovation activities.
In this paper we follow D’Este et al. (2008. 2009) and study the perception of innovation
barriers across two groups of non-innovators. We differ from D’Este et al. (2008) by using a
different filtering method of identifying barrier-related non-innovators based on the answers
to the questions on hampering factors to innovation as collected in CIS4 and CIS 2006. At the
same time we extend the research by D’Este et al. (2008) by considering a large number of
countries (18) and using a different set of control variables. Thus we are able to look more into
detail into the group of non-innovative firms and differentiating between different types of
innovative and non-innovative firms. By distinguishing barrier-related and non-barrier-related
non-innovative firms we are able to provide a richer picture on the barriers to innovation and
their importance across firms and countries.
2.3 Factors affecting the perception of innovation barriers
In our study we control for number of firm and sector characteristics. This is likely to improve
our understanding of attenuating obstacles to innovation. Based on the economic literature
on innovation we expect the following factors to affect the perception of deterring and/or
revealed barriers to innovation.
2.3.1 Firm size
Firm size is generally considered to be an important factor that explains firms’ innovation
behavior (e.g. Cohen and Klepper 1996). The ability of firms to create, access and
commercialize new knowledge, new financing and new skills is fundamental to sustained
growth at the firm level. Smaller firms in general have more difficulties than large firms to
afford the absorptive capacities to acquire knowledge and the means to access the
knowledge and to establish the collaborations required for their innovation activities. The fact
that larger firms are able to draw on an internal pool of resources and the fact that
innovative activities have some aspect of fixed outlays leads us to expect that larger firms are
less vulnerable to revealed barriers to innovation. Moreover is more likely that large noninnovative firms are non-innovative by choice, while for small firms barriers may deter them
from entering innovation activities.
2.3.2 Firm growth
Entrepreneurial high growth firms play an important role in radical innovation (Baumol 2007).
Moreover, there is evidence that new and small firms play a different role in competitive
markets (Acs and Audretsch 1987). In fact, innovation projects are often part of a growth
– 7 –
strategy. Firm growth requires new financial, human and knowledge resources. Hölzl (2009)
and Coad and Rao (2008) document that R&D and innovation are important determinants
high growth in advanced countries. There are reasons to think that high growth firms perceive
higher barriers to innovation because they were judge barriers lighter than other firms
because they were successful or not hampered by the factors. However, we expect that a
high growth performance does affect the perception of innovation barriers positively,
because such firms made much more effort to overcome the obstacles. We expect that high
growth firms rank innovation barriers higher than average firms and that firms with very low
growth rates do experience obstacles to innovation less than average firms.
2.3.3 Status of internationalization of the firm’s activities
There are a number of studies on the interaction between innovation and international
activities. The degree of internationalization of firms is often seen as result of the firm's
innovative activities (e.g. Markusen 1984, Rugman 1981). On average multinationals tend to
be larger, have a higher level of accumulated competence and tend to be more researchintensive than purely domestic firms (Iammarino et al. 2007). Technological activity in modern
multinationals is organized in international networks that allow the strategic integration of
different paths of innovation (e.g. Cantwell 1995, Veugelers and Cassiman 2004). The extent
to which multinational enterprises engage in innovative activities depend on their
technological strategy and the characteristics of the host environment. Thus being part of a
multinational group should reduce the perception that the lack of technological and market
knowledge acts as an innovation barrier. Knowledge and information transfer within
connected firms broadens the available knowledge base (know-what, know-how and knowwho). This should have an effect if the firm is part of a domestic corporate group. The effect
of internationalization in the form of exporting is more difficult to assess. The evidence that
internationalized firms operate internationally and are subject to competitive pressure from
firms from other countries leads to suspect that exporting firms are more aware of
technological knowledge gaps than firms that operate only domestically. Internationalized
firms access foreign markets that may follow slightly different customs and rules than the
domestic market. For this reason differentiated market knowledge is more relevant to them
than to domestic firms
3. Data and Method
3.1 Data sources
We use Community innovation Survey (CIS) data for 18 countries. In particular we use the CIS4 and CIS-2006 waves of the CIS. The Community Innovation Survey is a firm level survey
conducted every 4 years in all EU member states, as well as several non-EU countries (e.g.
– 8 –
Norway, Iceland).4 The CIS aims to provide a sound source of statistical data on innovation by
using a stratified sample of companies. CIS data are increasingly being used as a key data
source in the study of innovation at the firm level in Europe, Canada and Australia. Mairesse
and Mohnen (2004) provide evidence that the subjective measures of the CIS appear to be
consistent with objective measures of innovation, such as the probability of holding a patent
and the share in sales of products protected by patents.
3.2 Country groups
In this paper we apply implicitly the technology frontier concept at the country level. The
'distance to the frontier' approach takes into account the specific role of 'appropriate
institutions' (Aghion and Howitt 2006) at different stages of development. This approach
emphasizes that the effectiveness of economic policies is conditional to a country's distance
to the world technological frontier. While catch-up countries will profit more from capital
accumulation growth strategies, industrialized countries close to the technological frontier
are required to use an innovation-based strategy. Using a stylised model, Acemoglu et al.
(2006) show that high-skilled personnel and technology-intensive firms are more important to
economic growth in countries that are close to the technological frontier than for countries
further from the frontier.
We control for country differences by defining groups of countries that have approximately
the same position in technological development. Our classification of countries into different
groups is based on the research by Reinstaller and Unterlass (2010), who presented a
classification of EU countries based on the direct and indirect R&D intensity of each country
resulting from an input-output analysis. The direct R&D intensity is the direct investment of the
business sector into research and development as shown by the share of R&D in GDP of the
business sector in the common STI statistics. The indirect R&D intensity instead captures the
R&D embodied in capital goods used in the industries of a country. Together the two
indicators provide a rough measure of the level of technical development of a country in
terms of its capability to generate new technologies and its ability to use foreign
technologies. Reinstaller and Unterlass (2010) use cluster analysis to identify four country
groups: The first group of countries has high direct technology intensity and the relative share
of indirect technology intensity decreases with respect to other country groups. The countries
in the second group have high indirect technology intensity. Direct R&D intensity in these
countries is low, but R&D embodied in imported equipment is high. The countries in the third
group have relatively low levels of both direct and indirect technology intensity. The fourth
group, finally, consists of countries with low overall technology intensity both in terms of direct
and indirect R&D. Table 2 presents the classification of countries and indicates for which
countries CIS data could be accessed at the Eurostat Safe Centre in Luxemburg.
This data was accessed at the Safe Centre in Luxembourg. We wish to thank Sergiu Parvan at Eurostat. Without his
help this study would not have been possible.
4
– 9 –
Table 2: Country classification and data availability
Country group 1 (high direct technology intensity):
Belgium (BE)§, Denmark (DK)++,+++, Germany (DE)§, Finland (FI)++,+++, France (FR)++, Iceland (IS)++, Luxemburg
(LU)++,+++, Norway (NO)++,+++, Sweden (SE)++,+++, United Kingdom (UK)§, Netherlands (NL)§, Austria (AT)§
Country group 2 (high indirect technology intensity):
Czech Republic (CZ)++,+++, Estonia (EE)++,+++, Hungary (HU)++,+++, Slovenia (SI)++,+++, Slovak Republic (SK)++,+++,
Ireland (IE)+++
Country group 3 (low direct and indirect technology intensity, with higher GDP per capita):
Spain (ES)++,+++, Italy (IT)++,+++, Portugal (PT)++,+++, Greece (GR)++,+++
Country group 4 (low overall technology intensity):
Bulgaria (BG)++,+++, Lithuania (LT)++,+++, Latvia (LV)++,+++, Poland (PL)§, Romania (RO)++,+++, Cyprus (CY)+++, Malta
(MT)+++
Notes: Availability of Community Innovation Survey (CIS) data at the Eurostat Safe Centre in Luxemburg:
CIS2006; § access not allowed by national statistical institute.
++
CIS 4,
+++
3.2 Types of innovators and non-innovators
3.2.1 Innovators
We distinguish between innovating firms and non-innovating firms in a first step. We define all
firms that introduced a new or significantly improved product or process and/or have
ongoing innovation projects as innovators. In order to reduce the heterogeneity within the
group of innovators we distinguish two types of innovators: R&D innovators is the set of
innovative firms, that perform own R&D. The set of innovators that do not perform own R&D is
called non-technological innovators. The reason for this distinction is the fact that in
comparison to non-technological innovation related to R&D activities is generally more costly
and uncertain. This is likely to lead to a selection problem, when analyzing barriers to
innovation. We expect that R&D innovators have a different perception of obstacles to
innovation than non-technological innovators.
3.2.2 Procedure to identify barrier-related non-innovators
A similar selection problem arises for non-innovators. Here the puzzle arises that there is a
positive relationship between innovation barriers and the likelihood to be an innovator. This
correlation is a recurrent problem in the study of obstacles to innovation using the Community
Innovation Survey (e.g. Mohnen and Röller 2005, Iammarino et al. 2009, Lööf and Heshmati
2006). The existing literature on barriers to innovation has concentrated on the perception of
barriers among innovative firms (e.g. Mohnen and Rosa 1999, Baldwin and Lin 2002) or
treated non-innovative firms as an undifferentiated group (Hölzl and Friesenbichler 2009,
– 10 –
Iammarino et al. 2009). Therefore, most papers on innovation barriers do not take into
account their impact on the propensity to innovate (e.g. Canepa and Stoneman 2009, Galia
and Legros 2004). Moreover, this does not address the core of the policy question which
barriers to innovation are crucial in inhibiting the take-up of innovation activities by noninnovators
For this reason we follow D'Este et al. (2008, 2009) and distinguish explicitly between noninnovators that are rather indifferent about innovation activities and those that have some
aspiration to be innovative. Using this differentiation D'Este et al. (2008) are able to show that
non-innovators that have not much interest in performing innovation activities rank obstacles
to innovation very low. Savignac (2008) and Mancusi and Verzulli (2009) are able to show that
once firms not interested in innovation are excluded from the sample the positive correlation
vanishes and becomes negative. The sample selection bias arises because non-innovative
firms that do not aspire to be innovative do not meet any obstacle to innovative activities,
while firms that wished to innovate experience deterring barriers to innovation. By
distinguishing two groups of non-innovators D’Este et al. (2008) were able to identify deterring
barriers to innovation that are distinct from the revealed obstacles to innovation perceived
by innovating firms. The first are barriers that obstruct firms' achievement in innovation
activates, the second type of barriers prevents firms from engaging in innovation activities.
The distinction between the different groups of non-innovators is crucial for the present study.
Unfortunately the questions used by D'Este (2008) to distinguish between different types of
innovators is unique to the UK CIS-4 but are not available for CIS-4 and the CIS-2006 in the
harmonized sample we use. We will therefore use a different approach to identify innovationinterested from non-interested non-innovators. Specifically we use the intensity of answering
"high" or "medium" to all different specific barriers mentioned in the question "During the years
2004 to 2006, how important were the following factors for hampering your innovation
activities or projects or influencing a decision not to innovate" in the CIS-4/CIS-2006 as starting
point.
The starting point for our distinction is the assumption that non-barrier-related non-innovators
are firms that do not aspire to perform innovation activities. In order to distinguish between
barrier-related and non-barrier-related non-innovators we follow the following identification
scheme:
We define an indicator for barrier-relatedness for each firm (innovators and non-innovators),
that we define as average over the 9 different answers on the barriers (3=high, 2=medium,
1=low, 0=not experienced).
In order to control for the variety of answers across sectors and countries we subtract sectorcountry averages from the indicator for barrier-relatedness at the firm level. In addition, we
give those firms that show a higher variety of answers a higher weight by multiplying the
indicator of the barrier-relatedness by 1 + the standard deviation of answers to the 9
questions at the firm-level.
– 11 –
We calculate the average of the indicator for barrier-relatedness over the whole sample and
define as barrier-related innovators those non-innovating firms that have an above average
barrier-relatedness and gave low ranking to two questions that capture the reasons not to
innovate (“No need due to prior innovations” and “No need because of no demand for
innovations”). The other non-innovators are classified as non-barrier-related non-innovators
Figure 1: Distribution of innovator types across country groups
70%
60%
50%
RD innvovators
40%
non-technology innovators
30%
barrier-related noninnovators
20%
non-barrier-related noninnovators
10%
0%
all
country country country country
group 1 group 2 group 3 group 4
Source: CIS-4 and CIS-2006 data accessed at Eurostat Safe Centre; WIFO calculations. Values are averages over CIS4 and CIS-2006 aggregates. Country group 1: Belgium, Denmark, Germany, Finland, France, Iceland, Luxemburg,
Norway, Sweden; Country group 2: Czech Republic, Estonia, Hungary, Slovenia, Slovak Republic, Ireland; Country
group 3: Spain, Italy, Portugal, Greece; Country group 4: Bulgaria, Lithuania, Latvia, Romania, Cyprus, Malta.
Figure 1 presents the distribution of the types of innovators (R&D innovators, nontechnological innovators) and non-innovators (barrier-related and non-barrier-related noninnovators) across the country groups. Non-barrier-related non-innovators constitute the
largest group in all country groups. Country group 1 has the highest number of R&D
innovators, followed by country groups 2 and 3. Country group 4 has the lowest number of
R&D innovators. The opposite is true for non-barrier-related non-innovators. The distribution of
non-technology innovators is much more similar across the country groups. The highest share
is in country group 3 followed by country groups 3, 2 and 1. In contrast the distribution of
barrier-related non-innovators across country groups is quite unequal. Most barrier-related
non-innovators are found in country group 4 followed by country group 2 and 3. Country
group 1 has the lowest share of barrier-related non-innovators.
– 12 –
3.3 Barriers to innovation
The community innovation survey has information on nine different potential barriers to
innovation. In this study we consider five different barriers to innovation (in brackets the
original wording of the CIS questionnaire if different):
(i)
(ii)
(iii)
(iv)
(v)
financial barriers to innovation (Lack of finance from sources outside your
enterprise),
skill barriers to innovation (Lack of qualified personnel),
lack of information on technology,
lack of information on markets, and
lack of innovation partners (Difficulty in finding cooperation partners for
innovation).
Firms are asked to assess the importance of these barriers using a 4 valued scale from high
importance over medium to low importance and not relevant. From these answers we
construct a binary variable that takes on the value of 1 if the firm considers the degree of
importance of the barrier as high or medium. The variable takes on the value of 0 if the firm
considers the barrier of low importance or not relevant at all. The rationale for constructing
the dependent variable in this way is that we have then a indicator that discriminates
whether firms judge the barrier to be important or not.5
Table 3 provides descriptive evidence on differences in the perception of innovation barriers
across types of innovators and country groups. From these descriptive statistics emerges
clearly that skill barriers to innovation is the single most mentioned barrier to innovation,
followed by lack of external financing and knowledge barriers to innovation. Across country
groups the pattern emerges the relevance of barriers is increasing with technological
distance. Firms in country group 1 report in general the lowest number, followed by country
group 2, county group 3 and last country group 4.
The definition of barrier-related non-innovators used in this study implies that this type of firm
experiences barriers highest, especially the lack of financing, followed by R&D innovators and
non-technological innovators. Non-barrier-related non-innovators record the lowest scores for
innovation barriers. These firms are in general not hampered by their innovation activities
because they do not aspire to engage in innovation.
5 The reduction of informational content of the dependent variable – we do not differentiate between high and
medium on the one hand and low and not relevant on the other hand – allows us to use probit regression models
instead of ordered models that would take into account all four characteristics of the original variables. Specification
tests have shown that ordered probit failed to converge in a number of specifications and that no qualitative
differences with regard to interpretation emerge.
– 13 –
Table 3: Importance of selected barriers to innovation for all firms and innovators across country
groups
All
Country group 1 Country group 2 Country group 3 Country group 4
All firms
Financial constraints
33%
19%
28%
38%
Skill constraints
36%
34%
29%
37%
42%
39%
Lack of technological
knowledge
28%
19%
18%
33%
30%
Lack of market knowledge
27%
20%
19%
29%
28%
Lack of innovation
knowledge
25%
19%
20%
27%
33%
Financial constraints
44%
30%
55%
55%
R&D innovators
38%
Skill constraints
47%
49%
47%
45%
54%
Lack of technological
knowledge
33%
28%
25%
37%
35%
Lack of market knowledge
33%
31%
28%
34%
35%
Lack of innovation
knowledge
32%
28%
25%
37%
37%
Financial constraints
38%
19%
43%
48%
Non-technological innovators
30%
Skill constraints
42%
42%
35%
42%
45%
Lack of technological
knowledge
34%
23%
19%
38%
32%
Lack of market knowledge
30%
22%
20%
32%
30%
Lack of innovation
knowledge
26%
20%
20%
27%
36%
Financial constraints
62%
37%
69%
73%
Barrier-related non-innovators
57%
Skill constraints
61%
59%
49%
64%
60%
Lack of technological
knowledge
49%
33%
32%
57%
48%
Lack of market knowledge
47%
33%
33%
52%
45%
Lack of innovation
knowledge
45%
34%
37%
48%
52%
Non-barrier-related non-innovators
Financial constraints
20%
9%
17%
22%
27%
Skill constraints
24%
17%
16%
25%
28%
Lack of technological
knowledge
20%
10%
11%
23%
22%
Lack of market knowledge
18%
9%
12%
21%
21%
Lack of innovation
knowledge
17%
10%
14%
18%
24%
Source: CIS 4 and CIS 2006 data accessed at the Eurostat safe centre. WIFO calculations. The numbers are simple
averages over CIS-4 and CIS-2006 averages. See section 2.1 for details on the country groups (group 1 : member
states close to technological frontier, group 2 : advanced catching up member states, group 3 : Southern European
member states with low- to medium tech industry structure and high GDP, group 4 : trailing catching up member
– 14 –
states). Barriers are measured as binary variable. The variable takes the value of 1 if the degree of importance is
judged to be medium or high. If the degree of importance is judge to be low or not relevant the variable gets the
value 0. Country group 1: Belgium, Denmark, Germany, Finland, France, Iceland, Luxemburg, Norway, Sweden;
Country group 2: Czech Republic, Estonia, Hungary, Slovenia, Slovak Republic, Ireland; Country group 3: Spain, Italy,
Portugal, Greece; Country group 4: Bulgaria, Lithuania, Latvia, Romania, Cyprus, Malta.
The lack of qualified personnel is ranked higher than the lack of financial resources in country
group 1 and country group 2 for innovating firms. Financial resources are considered to the
more important barrier by firms in the other country groups, especially in country group 4. The
knowledge barriers related to lack of knowledge on technology, lack of knowledge on
markets and difficulties of finding innovation partners are ranked higher in country groups 3
and 4 than in country groups 1 and 2. The differences between country groups are not only
found at the aggregate level but also at the level of the different country groups. For
example, the higher importance of skill constraints for firms in country group 1 is found for R&D
innovators but also for barrier-related and non-barrier-related non-innovators.
The overall message is that differences matter for the perception and experience of barriers
to innovation:
1. The perception of barriers to innovation is higher for innovating firms than for noninnovating firms. R&D innovators perceive barriers to innovation as more important
than nontechnology innovators. This clearly shows that hampering factors should be
considered as barriers that can be overcome, at least by innovative firms.
2. The distance to the frontier matters for the perception and experience of different
innovation barriers. Firms in countries closer to technological frontier attach more
importance to the lack of skilled labour than to lack of financing. For countries far
away from the distance from the frontier it is the opposite.
Thus we will carry out an econometric analysis for the different country groups, in order to
substantiate this impression.
3.4 Variable definitions
The definition of the barriers to innovation variables, the country groups and the innovation
types was presented before. For the econometric analysis we use as mentioned in section 2.3
we are also interested in factors affecting the perception of innovation barriers. We use the
following variables:
1. Firm size is measured by the logarithm of employees. As discussed earlier firm size is
generally considered to be an important factor affecting the propensity to innovate and
therefore provides evidence on the issue whether smaller or larger firms perceive the
– 15 –
specific barriers more importantly. It is generally assumed that because of their small size
SMEs experience higher barriers to innovation.
2. The second set of variables is related to the growth performance of enterprises. We define
high growth firms following a definition that is oriented at the Eurostat-OECD definition. For
each enterprise with more than 10 employees we calculate the annualized employment
growth rate:
growth
E
E
,
,
-1.
High growth firms are those firms whose annualized growth rate is above 20%. Given the
fact that high growth firms are important for net employment generation and the
diffusion of technology it is surprising that we do not know more about them (e.g.
Henrekson and Johansson 2010, Coad and Hölzl 2010). In order to provide a contrast to
the high growth firms and to be able to interpret the effect of growth on the perception
of innovation barriers we select a set of firms that did have only marginal growth rates in
the range of -3% and + 3% annualized growth. We call these firms stable firms.
3. It is generally thought that it is easier for firms to access knowledge sources and human
and financial ressources within the same enterprise than going outside of the enterprise.
Knowledge and financial resources can be shared between the firms of an enterprise
(internal markets for know-how and internal financial markets). In addition critical human
capital can be can be temporarily reallocated between enterprises belonging to the
same corporate group much more easily than between independent firms. For that
reason we include dummy variables that identify whether the firm is part of a corporate
group. We take foreign ownership separately into account, as foreign ownership may
give access to better technology than is possible within domestic company groups. In
fact, the literature on multinationals emphasises that multinationals tend to be larger,
have a higher level of accumulated competence and tend to be more researchintensive than purely domestic firms (Cantwell 1995, Iammarino et al. 2007). Thus
belonging to an enterprise group may affect the importance and perception of
obstacles. Thus we use separate dummy variables indicating whether a firm is part of a
foreign multinational, and wether a firm is part of a domestic corporate group. In
addition, following the literature on exporting and innovation, we include also a dummy
variable for internationalisation into the analysis.
4. In addition we employ a number of control variables that relate to the sectors. We use a
dummy variable indicating whether the firm operates in the manufacturing sector or not
(manuf). In addition we use aggregated industry classification dummies following the
– 16 –
innovation taxonomy by Peneder (2010) that distinguishes 5 different sector groups
according to innovation intensity in the country group regressions (see appendix A)..
4. Results
4.1 Basic methodology
The primary goal of the analysis is to uncover systematic differences between different types
of innovative and non-innovative firms and factors affecting the perception of innovation
barriers across country groups. This limits the construction of dependent variables to the
questions in the CIS that are answered by all firms. .
Our baseline specification is the following:
Barrier = f(FS, GAZ, STABLE, INTER, GP_fo, GP_do, INDUSTRY, COUNTRYGR, INNOTYPE)
Where FS denotes firm size, GAZ the fact whether the firm is a fast growing firm, STABLE the
fact whether the firm experienced low growth/decline, INTER whether the firm is
internationalised or not, GP_fo is a dummy variable denoting that the firm is part of a foreign
corporate group, GP_do is a dummy variable denoting that the firm is part of a domestic
corporate group. INDUSTRY denotes a set of dummy variables where manuf is a dummy
indicating that the firm is in the manufacturing sector. In addition we use aggregated industry
classification dummies following the innovation taxonomy by Peneder (2010) that
distinguishes 5 different sector groups according to innovation intensity in the country group
regressions. COUNTRYGR denotes the country group dummies and last but not least
INNOTYPE denotes the four different definitions of Innovators and non-innovators.
This specification is estimated using probit regressions. In the regression the weights provided
by Eurostat were used in order to correct for the different sampling of firms at the country
level.
(to be expanded)
4.2 Baseline Results
Table 5 reports the results of the baseline regressions. The results are quite similar between CIS
4 and CIS 2006, thus suggesting that we in fact uncover quite robust regularities in the
perception of innovation barriers across firms. The results show that firm size has the expected
negative influence on the perception of innovation barriers. Larger firms do not perceive
innovation barriers to be as strict as smaller firms. Gazelles judge innovation barriers to be
higher than average firms (this is the reference group) with the exception of lack of
– 17 –
knowledge on markets for CIS 2006. Stable firms in contrast, are hampered less by innovation
barriers.
Interestingly, export-active firms report higher innovation barriers than firms that operate only
in national markets (implicit reference group). This might be associated with the fact that
internationalized firms are those firms that self-select into exporting because they are more
productive and innovative. At the international markets they face tighter competition from
similar innovative firms. In contrast being part of a foreign or a domestic group reduces the
perception of innovation barriers considerably. The effect is much stronger for being part of a
foreign group than for a domestic group for the CIS 4 sample, but not for the CIS 2006
sample. This may be associated with the different coverage of countries in country group 1 in
the CIS 2006 sample.
Manufacturing firms report a higher impact of obstacles to innovation than nonmanufacturing firms (reference group). The dummy variables associated with the industry
innovation taxonomy of Peneder (2010) shows that firms in industries with high innovation
intensity report higher obstacles with regard to lack of innovation partners, financing barriers
and skill barriers but lower barriers with regard to lack of information about technology and
markets. These taxonomies were used primarily as control variables and should not be overinterpreted. The proxies for basicness and cumulativeness of the knowledge base of the
innovation process provide interesting and expected results. Basicness of the knowledge
base is generally associated with a higher level of innovation barriers with the exception of
skilled labour, while a more cumulative knowledge base is associated with a lower
perception of innovation barriers.
The results for the country groups shows that firms located in country group 1 and country
group 2 report cetaris paribus lower innovation barriers than firms in country group 4
(reference group). This holds also for country group 3 for the CIS 2006 but not for the CIS 4.
With regard to the innovator types we see that R&D innovators, non-technology innovators
and barrier-related non-innovators have a significantly higher propensity to assess innovation
barriers as relevant than non-barrier-related non-innovators. The ranking of enterprise types is
in general that barrier-related non-innovators have the highest propensity to be affected by
innovation barriers. The second group are the R&D innovators followed by non-technology
innovators. Non-barrier-related non-innovators (reference group) have the lowest propensity
to rank innovation barriers as relevant.
It is interesting to see that both R&D innovators and especially barrier-related non-innovators
rank all innovation barriers as high. In contrast non-technology innovators give a lower
ranking to the lack of innovation partners, thus reflecting the different worlds of R&D
innovation, non-technological innovation and deterring barriers to innovation.
– 18 –
Table 5: Results of the baseline regressions
CIS 4
CIS 2006
lack of
technical
knowledge
lack of
market
knowledge
lack of
innovation
partners
financial
barriers
-0.0089***
(-15.205)
gazelle (y/n)
0.0164***
(8.213)
stable (y/n)
-0.0074***
(-6.262)
exporting (y/n)
0.0049***
(3.937)
part of foreign group
-0.0516***
(-23.549)
part of domestic group
-0.0184***
(-11.489)
manuf
0.0525***
(29.738)
medium-low innovation
-0.0534***
(-17.009)
medium innovation
-0.0378***
(-15.288)
medium-high innovation
-0.0337***
(-8.499)
high innovation
-0.0440***
(-10.940)
country group 1
-0.0480***
(-20.757)
Country group 2
-0.0753***
(-28.132)
Country group 3
0.0481***
(21.863)
Basicness
0.0196
(1.277)
Cumulativeness
-0.1205***
(-16.485)
R&D innovator
0.1332***
(75.909)
non-technology innovator
0.1263***
(91.127)
barrier-related non-innovato 0.2608***
(171.945)
Constant
0.4978***
(30.100)
-0.0060***
(-10.443)
0.0087***
(4.405)
-0.0142***
(-12.201)
0.0048***
(3.940)
-0.0591***
(-27.367)
-0.0196***
(-12.416)
0.0581***
(33.354)
-0.0434***
(-14.000)
-0.0119***
(-4.884)
-0.0060
(-1.546)
-0.0026
(-0.658)
-0.0382***
(-16.762)
-0.0581***
(-21.985)
0.0227***
(10.455)
0.0117
(0.770)
-0.1151***
(-15.977)
0.1299***
(75.041)
0.0931***
(68.102)
0.2500***
(167.127)
0.4700***
(28.816)
-0.0072***
(-12.605)
0.0084***
(4.285)
-0.0164***
(-14.284)
0.0187***
(15.448)
-0.0535***
(-25.096)
-0.0088***
(-5.680)
0.0339***
(19.685)
-0.0194***
(-6.355)
-0.0023
(-0.970)
0.0171***
(4.436)
0.0185***
(4.736)
-0.0909***
(-40.345)
-0.0930***
(-35.676)
-0.0224***
(-10.467)
0.1051***
(7.011)
-0.1563***
(-21.959)
0.1301***
(76.118)
0.0545***
(40.421)
0.2598***
(175.868)
0.5876***
(36.474)
707,392
0.057
-390456
707,400
0.063
-381602
log firm size
Observations (weighted)
pseudo R2
ll
707,397
0.068
-400233
skill barriers
lack of
technical
knowledge
lack of
market
knowledge
lack of
innovation
partners
financial
barriers
skill barriers
-0.0150***
(-25.299)
0.0101***
(4.944)
-0.0284***
(-23.630)
0.0130***
(10.234)
-0.0727***
(-32.514)
-0.0067***
(-4.113)
0.0728***
(40.375)
-0.0426***
(-13.275)
-0.0212***
(-8.375)
-0.0213***
(-5.264)
0.0174***
(4.252)
-0.1659***
(-70.265)
-0.0960***
(-35.096)
-0.0001
(-0.026)
0.0678***
(4.319)
-0.1417***
(-18.997)
0.2037***
(113.725)
0.1319***
(93.261)
0.3436***
(221.862)
0.6333***
(37.510)
-0.0095***
(-15.187)
0.0258***
(12.008)
-0.0147***
(-11.609)
-0.0151***
(-11.337)
-0.0657***
(-27.950)
-0.0233***
(-13.568)
0.0603***
(31.777)
-0.0212***
(-6.298)
0.0092***
(3.463)
0.0262***
(6.156)
0.0512***
(11.868)
0.0032
(1.270)
-0.0663***
(-23.077)
0.0096***
(4.056)
-0.1611***
(-9.755)
-0.1325***
(-16.882)
0.1997***
(106.012)
0.1580***
(106.235)
0.3288***
(201.922)
0.6146***
(34.611)
-0.0070***
(-8.738)
0.0065**
(2.478)
0.0009
(0.519)
-0.0024
(-1.299)
-0.0546***
(-17.282)
-0.0500***
(-17.703)
0.0798***
(31.462)
-0.0371***
(-8.963)
-0.0151***
(-4.571)
-0.0172***
(-3.428)
-0.0263***
(-5.238)
-0.1588***
(-41.826)
-0.1979***
(-70.211)
-0.0348***
(-14.932)
-0.0244
(-1.093)
-0.1521***
(-13.182)
0.1646***
(56.417)
0.1273***
(62.484)
0.2974***
(133.159)
0.6662***
(25.852)
-0.0104***
(-13.128)
-0.0073***
(-2.820)
-0.0052***
(-2.972)
0.0105***
(5.805)
-0.0489***
(-15.794)
-0.0523***
(-18.872)
0.0712***
(28.577)
-0.0457***
(-11.259)
-0.0192***
(-5.944)
-0.0356***
(-7.217)
-0.0178***
(-3.611)
-0.1469***
(-39.449)
-0.1799***
(-65.051)
-0.0506***
(-22.123)
0.1419***
(6.465)
-0.2082***
(-18.390)
0.1737***
(60.648)
0.1144***
(57.208)
0.2838***
(129.498)
0.7633***
(30.181)
-0.0027***
(-3.517)
0.0056**
(2.222)
-0.0165***
(-9.593)
0.0185***
(10.427)
-0.0619***
(-20.424)
-0.0243***
(-8.966)
0.0559***
(22.935)
-0.0129***
(-3.245)
0.0155***
(4.914)
0.0151***
(3.128)
0.0233***
(4.842)
-0.1461***
(-40.074)
-0.1544***
(-57.016)
-0.0819***
(-36.553)
0.0794***
(3.698)
-0.1881***
(-16.973)
0.1389***
(49.553)
0.0767***
(39.206)
0.2621***
(122.180)
0.6985***
(28.216)
-0.0102***
(-12.615)
0.0132***
(5.003)
-0.0139***
(-7.722)
0.0266***
(14.347)
-0.1169***
(-36.883)
-0.0432***
(-15.240)
0.0873***
(34.283)
-0.0077*
(-1.861)
-0.0152***
(-4.598)
-0.0106**
(-2.107)
0.0078
(1.542)
-0.1744***
(-45.777)
-0.1164***
(-41.143)
-0.0212***
(-9.058)
0.2454***
(10.929)
-0.2168***
(-18.717)
0.2374***
(81.038)
0.1604***
(78.439)
0.4319***
(192.665)
0.7513***
(29.043)
-0.0046***
(-5.518)
0.0118***
(4.289)
-0.0069***
(-3.705)
0.0053***
(2.766)
-0.0684***
(-20.785)
-0.0497***
(-16.862)
0.0998***
(37.713)
-0.0278***
(-6.445)
-0.0063*
(-1.842)
-0.0198***
(-3.787)
0.0209***
(4.000)
-0.0901***
(-22.775)
-0.1576***
(-53.615)
-0.0741***
(-30.464)
-0.0247
(-1.061)
-0.1277***
(-10.613)
0.2316***
(76.114)
0.1719***
(80.961)
0.3681***
(158.083)
0.6474***
(24.092)
707,373
0.120
-414859
707,434
0.075
-450575
340,636
0.087
-203617
340,640
0.080
-197222
340,646
0.068
-189942
340,677
0.142
-204969
340,652
0.103
-217856
Source: CIS 4 and 2006 data accessed at the safe centre. t-statistics in parentheses, *** p<0.01, ** p<0.05, * p<0.1.
Country group 1: Belgium, Denmark, Germany, Finland, France, Iceland, Luxemburg, Norway, Sweden; Country
group 2: Czech Republic, Estonia, Hungary, Slovenia, Slovak Republic, Ireland; Country group 3: Spain, Italy, Portugal,
Greece; Country group 4: Bulgaria, Lithuania, Latvia, Romania, Cyprus, Malta.
4.3 Differences between country groups
In order to go one step deeper we estimated the baseline regressions for the country groups
separately.6 As it is impossible to extrapolate the statistical significance of the differences
between country groups we explicitly test for these differences. Table 6 presents the
differences of coefficients between country groups. In order to test for the equality of
coefficients, we run a OLS regression using the sample of two country groups. For one of the
6
The results of the country group regressions are in appendix B.
– 19 –
country groups the coefficients were interacted with a dummy variable. The regression results
then reported the OLS results for one country group (the country group with the lower
number) while the interacted results showed the difference of coefficient between the
country groups. The reading of table 6 is the following: A positive expression indicates the
coefficient for the second country group is higher, while a negative expression indicates that
the coefficient for the first country group is higher. The regressions run were the same as the
baseline regressions except for the exclusion of country group dummies. The table does
report only factors affecting the perception of barriers and innovator types.
Table 6: Differences of coefficients between country groups using country group regressions,
CIS 4
CG 1 vs CG 2
lack of technical knowledge
gazelle (y/n)
0,046
stable (y/n)
-0,002
exporting (y/n
0,015
part of foreign
-0,031
part of domes
0,050
R&D innovato
0,024
non-technolog
-0,082
barrier-related
-0,043
6,1
-0,7
3,7
-4,9
7,9
3,9
-16,0
-8,7
CG 1 vs CG 3
0,019
0,020
0,007
-0,101
-0,023
-0,013
-0,018
0,056
4,0
7,4
2,6
-19,9
-6,8
-3,3
-5,5
16,2
CG1 vs CG 4
0,038
0,030
0,008
-0,026
0,014
-0,020
-0,036
-0,003
5,3
4,5
1,4
-2,3
1,4
-1,1
-5,2
-0,5
CG 2 vs CG 3
-0,028
0,022
-0,008
-0,070
-0,073
-0,037
0,064
0,099
-3,3
5,3
-1,7
-9,0
-10,0
-5,5
11,7
18,7
CG 2 vs CG 4
-0,008
0,032
-0,007
0,005
-0,036
-0,044
0,046
0,040
-0,9
4,5
-1,0
0,4
-3,1
-2,4
5,8
5,3
CG 3 vs CG 4
0,020
0,010
0,001
0,075
0,037
-0,007
-0,018
-0,059
2,5
1,4
0,2
5,8
3,2
-0,4
-2,3
-8,1
lack of market knowledge
gazelle (y/n)
0,045
stable (y/n)
-0,024
exporting (y/n
0,039
part of foreign
-0,039
part of domes
0,018
R&D innovato
0,022
non-technolog
-0,047
barrier-related
-0,054
5,9
-6,4
9,6
-6,1
2,8
3,7
-9,1
-11,0
0,006
-0,013
0,006
-0,096
-0,044
-0,079
-0,052
-0,023
1,4
-4,9
2,2
-19,2
-13,1
-20,5
-16,2
-6,7
0,019
0,004
0,033
-0,010
-0,012
-0,024
-0,015
-0,052
2,6
0,7
5,5
-0,9
-1,2
-1,4
-2,1
-7,9
-0,039
0,011
-0,033
-0,057
-0,062
-0,101
-0,006
0,032
-4,8
2,8
-7,6
-7,5
-8,7
-15,4
-1,0
6,1
-0,026
0,029
-0,006
0,029
-0,030
-0,046
0,032
0,003
-2,8
4,0
-0,9
2,3
-2,5
-2,5
4,0
0,4
0,012
0,017
0,027
0,086
0,033
0,055
0,038
-0,029
1,6
2,4
4,1
6,8
2,9
2,8
5,0
-4,0
lack of innovation partners
gazelle (y/n)
0,014
stable (y/n)
-0,038
exporting (y/n
0,025
part of foreign
-0,034
part of domes
0,018
R&D innovato
-0,037
non-technolog
-0,041
barrier-related
-0,024
1,9
-10,1
6,2
-5,3
2,8
-6,2
-8,0
-5,0
-0,001
-0,036
0,015
-0,033
0,005
-0,040
-0,077
0,018
-0,1
-14,1
5,4
-6,7
1,5
-10,5
-24,2
5,5
-0,002
-0,003
0,019
-0,044
-0,041
-0,056
-0,008
0,010
-0,3
-0,5
3,1
-4,0
-4,1
-3,3
-1,2
1,5
-0,015
0,002
-0,010
0,001
-0,013
-0,003
-0,036
0,043
-1,9
0,5
-2,4
0,1
-1,8
-0,4
-6,7
8,3
-0,017
0,035
-0,007
-0,010
-0,059
-0,020
0,033
0,034
-1,7
4,7
-1,0
-0,8
-4,9
-1,0
4,0
4,4
-0,002
0,033
0,004
-0,011
-0,046
-0,017
0,069
-0,008
-0,2
4,6
0,5
-0,8
-4,1
-0,9
9,2
-1,2
financial barriers
gazelle (y/n)
stable (y/n)
exporting (y/n
part of foreign
part of domes
R&D innovato
non-technolog
barrier-related
0,008
-0,020
0,026
-0,118
-0,011
0,055
0,045
0,114
1,0
-5,2
6,3
-18,4
-1,8
9,0
8,7
22,9
0,010
-0,019
-0,001
-0,124
-0,069
0,044
0,054
0,113
2,1
-7,2
-0,3
-24,2
-19,9
11,2
16,4
32,1
-0,008
-0,003
0,026
-0,147
-0,089
0,071
0,072
0,152
-1,1
-0,4
4,3
-13,3
-8,9
4,1
10,4
23,3
0,002
0,000
-0,027
-0,006
-0,057
-0,011
0,009
-0,001
0,3
0,1
-5,9
-0,7
-7,6
-1,6
1,6
-0,2
-0,015
0,017
0,000
-0,029
-0,078
0,016
0,027
0,038
-1,5
2,2
-0,1
-2,1
-6,1
0,8
3,1
4,7
-0,017
0,017
0,027
-0,023
-0,021
0,027
0,018
0,039
-2,2
2,2
3,8
-1,7
-1,7
1,3
2,3
5,3
skill barriers
gazelle (y/n)
stable (y/n)
exporting (y/n
part of foreign
part of domes
R&D innovato
non-technolog
barrier-related
0,002
-0,007
0,048
0,001
0,039
0,035
-0,092
-0,098
0,2
-1,5
10,1
0,1
5,4
5,0
-15,6
-17,4
-0,005
0,007
0,020
-0,017
-0,022
-0,082
-0,090
-0,075
-1,1
2,5
6,4
-3,1
-6,1
-19,5
-25,7
-20,1
0,020
0,019
0,033
-0,012
0,052
0,044
-0,048
-0,093
2,4
2,5
4,8
-0,9
4,5
2,2
-6,1
-12,3
-0,007
0,014
-0,028
-0,018
-0,061
-0,116
0,002
0,024
-0,8
3,1
-6,0
-2,2
-8,1
-16,6
0,3
4,3
0,018
0,025
-0,014
-0,012
0,013
0,010
0,044
0,006
1,7
3,2
-1,9
-0,9
1,0
0,5
4,9
0,7
0,025
0,012
0,014
0,005
0,075
0,126
0,042
-0,018
3,1
1,5
1,9
0,4
6,2
6,1
5,2
-2,4
Source: CIS 4 data accessed at the safe centre. Coefficients and t-statistics reported. The sign of the coefficient
indicates the country group which has the higher (more positive) coefficient, as it is the difference between the two
coefficients. A negative relationship indicates that the first country, e.g. in the first column country group 1, has the
higher coefficient. A positive coefficient indicates that the second country, e.g. in the first column country group 2,
has the higher coefficient. Country group 1: Belgium, Denmark, Germany, Finland, France, Iceland, Luxemburg,
– 20 –
Norway, Sweden; Country group 2: Czech Republic, Estonia, Hungary, Slovenia, Slovak Republic, Ireland; Country
group 3: Spain, Italy, Portugal, Greece; Country group 4: Bulgaria, Lithuania, Latvia, Romania, Cyprus, Malta.
The results in tables 6 and 7 provide some interesting insights regarding the factors affecting
the perception of innovation barriers across country groups:
1. High growth firms in country group 1 do report in general that they were affected less
by innovation barriers than high growth firms in the other country groups. In addition.
high growth firms in country groups 2 and 4 report in general to be more affected
than high growth firms in country group 3.
2. For stable firms (annualized growth rate in the interval -3% and +3%) the patterns are
not that clear cut. In fact, stable firms in country group 1 have a higher coefficient for
“lack of market information” and financial barriers but have a lower cordfficient for
skill constraints and lack of information on technology compared to stable firms in the
other country groups. For the other country groups the ranking is generally the
following CG4=>CG3>=CG2
3. Export active firms in country group 1 do report in general that they are affected less
by innovation barriers than export active firms in other country groups. The ranking for
the other country groups follows largely the following pattern CG2>= CG4>=CG3.
4. With regard to the dummy variable that the firm is part of a enterprise group we
observe an interesting pattern for country group 1, namely that the coefficient for
firms that are part of a foreign group is higher than for the other country groups while
this is not necessarily the case for domestic groups.
5. For the innovation types we are able to report on interesting result for country groups.
It appears that with the exception for financial barriers that innovation barriers affect
non-technology innovators and barrier related innovators in country group 1 more
than in the other country groups. Financial barriers are more relevant for firms in the
other country groups while for skill constraints and innovation partners the largest
differences are reported for country group 1. This holds especially for nontechnological innovators. With regard to R&D innovators alone R&D innovators in
country group 2 seems to be affected most by innovation barriers, with the exception
of lack of innovation partners.
(to be completed)
Table 7: Differences of coefficients between country groups using country group regressions,
CIS 2006
– 21 –
CG 1 vs CG 2
lack of technical knowledge
gazelle (y/n)
0,046
stable (y/n)
-0,002
exporting (y/n
0,015
part of foreign
-0,031
part of domes
0,050
R&D innovato
0,024
non-technolog
-0,082
barrier-related
-0,043
CG 1 vs CG 3
CG1 vs CG 4
CG 2 vs CG 3
CG 2 vs CG 4
CG 3 vs CG 4
6,1
-0,7
3,7
-4,9
7,9
3,9
-16,0
-8,7
0,019
0,020
0,007
-0,101
-0,023
-0,013
-0,018
0,056
4,0
7,4
2,6
-19,9
-6,8
-3,3
-5,5
16,2
0,038
0,030
0,008
-0,026
0,014
-0,020
-0,036
-0,003
5,3
4,5
1,4
-2,3
1,4
-1,1
-5,2
-0,5
-0,028
0,022
-0,008
-0,070
-0,073
-0,037
0,064
0,099
-3,3
5,3
-1,7
-9,0
-10,0
-5,5
11,7
18,7
-0,008
0,032
-0,007
0,005
-0,036
-0,044
0,046
0,040
-0,9
4,5
-1,0
0,4
-3,1
-2,4
5,8
5,3
0,020
0,010
0,001
0,075
0,037
-0,007
-0,018
-0,059
2,5
1,4
0,2
5,8
3,2
-0,4
-2,3
-8,1
lack of market knowledge
gazelle (y/n)
0,045
stable (y/n)
-0,024
exporting (y/n
0,039
part of foreign
-0,039
part of domes
0,018
R&D innovato
0,022
non-technolog
-0,047
barrier-related
-0,054
5,9
-6,4
9,6
-6,1
2,8
3,7
-9,1
-11,0
0,006
-0,013
0,006
-0,096
-0,044
-0,079
-0,052
-0,023
1,4
-4,9
2,2
-19,2
-13,1
-20,5
-16,2
-6,7
0,019
0,004
0,033
-0,010
-0,012
-0,024
-0,015
-0,052
2,6
0,7
5,5
-0,9
-1,2
-1,4
-2,1
-7,9
-0,039
0,011
-0,033
-0,057
-0,062
-0,101
-0,006
0,032
-4,8
2,8
-7,6
-7,5
-8,7
-15,4
-1,0
6,1
-0,026
0,029
-0,006
0,029
-0,030
-0,046
0,032
0,003
-2,8
4,0
-0,9
2,3
-2,5
-2,5
4,0
0,4
0,012
0,017
0,027
0,086
0,033
0,055
0,038
-0,029
1,6
2,4
4,1
6,8
2,9
2,8
5,0
-4,0
lack of innovation partners
gazelle (y/n)
0,014
stable (y/n)
-0,038
exporting (y/n
0,025
part of foreign
-0,034
part of domes
0,018
R&D innovato
-0,037
non-technolog
-0,041
barrier-related
-0,024
1,9
-10,1
6,2
-5,3
2,8
-6,2
-8,0
-5,0
-0,001
-0,036
0,015
-0,033
0,005
-0,040
-0,077
0,018
-0,1
-14,1
5,4
-6,7
1,5
-10,5
-24,2
5,5
-0,002
-0,003
0,019
-0,044
-0,041
-0,056
-0,008
0,010
-0,3
-0,5
3,1
-4,0
-4,1
-3,3
-1,2
1,5
-0,015
0,002
-0,010
0,001
-0,013
-0,003
-0,036
0,043
-1,9
0,5
-2,4
0,1
-1,8
-0,4
-6,7
8,3
-0,017
0,035
-0,007
-0,010
-0,059
-0,020
0,033
0,034
-1,7
4,7
-1,0
-0,8
-4,9
-1,0
4,0
4,4
-0,002
0,033
0,004
-0,011
-0,046
-0,017
0,069
-0,008
-0,2
4,6
0,5
-0,8
-4,1
-0,9
9,2
-1,2
financial barriers
gazelle (y/n)
stable (y/n)
exporting (y/n
part of foreign
part of domes
R&D innovato
non-technolog
barrier-related
0,008
-0,020
0,026
-0,118
-0,011
0,055
0,045
0,114
1,0
-5,2
6,3
-18,4
-1,8
9,0
8,7
22,9
0,010
-0,019
-0,001
-0,124
-0,069
0,044
0,054
0,113
2,1
-7,2
-0,3
-24,2
-19,9
11,2
16,4
32,1
-0,008
-0,003
0,026
-0,147
-0,089
0,071
0,072
0,152
-1,1
-0,4
4,3
-13,3
-8,9
4,1
10,4
23,3
0,002
0,000
-0,027
-0,006
-0,057
-0,011
0,009
-0,001
0,3
0,1
-5,9
-0,7
-7,6
-1,6
1,6
-0,2
-0,015
0,017
0,000
-0,029
-0,078
0,016
0,027
0,038
-1,5
2,2
-0,1
-2,1
-6,1
0,8
3,1
4,7
-0,017
0,017
0,027
-0,023
-0,021
0,027
0,018
0,039
-2,2
2,2
3,8
-1,7
-1,7
1,3
2,3
5,3
skill barriers
gazelle (y/n)
stable (y/n)
exporting (y/n
part of foreign
part of domes
R&D innovato
non-technolog
barrier-related
0,002
-0,007
0,048
0,001
0,039
0,035
-0,092
-0,098
0,2
-1,5
10,1
0,1
5,4
5,0
-15,6
-17,4
-0,005
0,007
0,020
-0,017
-0,022
-0,082
-0,090
-0,075
-1,1
2,5
6,4
-3,1
-6,1
-19,5
-25,7
-20,1
0,020
0,019
0,033
-0,012
0,052
0,044
-0,048
-0,093
2,4
2,5
4,8
-0,9
4,5
2,2
-6,1
-12,3
-0,007
0,014
-0,028
-0,018
-0,061
-0,116
0,002
0,024
-0,8
3,1
-6,0
-2,2
-8,1
-16,6
0,3
4,3
0,018
0,025
-0,014
-0,012
0,013
0,010
0,044
0,006
1,7
3,2
-1,9
-0,9
1,0
0,5
4,9
0,7
0,025
0,012
0,014
0,005
0,075
0,126
0,042
-0,018
3,1
1,5
1,9
0,4
6,2
6,1
5,2
-2,4
Source: CIS 4 data accessed at the safe centre. Coefficients and t-statistics reported. The sign of the coefficient
indicates the country group which has the higher (more positive) coefficient, as it is the difference between the two
coefficients. A negative relationship indicates that the first country, e.g. in the first column country group 1, has the
higher coefficient. A positive coefficient indicates that the second country, e.g. in the first column country group 2,
has the higher coefficient. Country group 1: Belgium, Denmark, Germany, Finland, France, Iceland, Luxemburg,
Norway, Sweden; Country group 2: Czech Republic, Estonia, Hungary, Slovenia, Slovak Republic, Ireland; Country
group 3: Spain, Italy, Portugal, Greece; Country group 4: Bulgaria, Lithuania, Latvia, Romania, Cyprus, Malta.
These results show how institutional and economic differences may shape innovation
activities. Let us consider financing constraints and skill barriers in more detail. Figure 2 shows
financial market development across the EU. The UK, Benelux and Scandinavian countries as
well as Spain feature the most developed markets, with the big continental economies
Germany, Italy and France following behind. The Eastern European countries are generally
least developed. Financial constraints bite particularly in countries with less developed
financial markets – emerging market countries -, where they not only hold back firms' growth
(Angelini and Generale 2008) but also prevent domestic firms from reaping the benefits from
trade liberalisation in terms of productivity gains (Gorodnichenko and Schnitzer 2010)
because innovative activities are constrained by the availability of external finance.
– 22 –
Figure 2 Financial market development (sum of stock market capitalisation, private credit and public
bonds/GDP), 2004-2006
(.86 520 84,1]
(.66 372 08,.86 520 84]
(.59 868 97,.66 372 08]
[0,.598 689 7]
Source: Beck et al. 2010, WIFO calculations.
Our results implicitly suggest that financial development reduces financing constraints to
innovation, but does not help in closing the skill constraints in the most advanced economies.
[to be developed]
– 23 –
4.4 Robustness
Table 8 presents robustness results that can be compared to the baseline results in table 5 in
order to assess the robustness of the results with regard to changes in the definition of barrierrelated non-innovators. We report the results for three changes in definition:
1. We exclude all barrier-related non-innovators that ranked all obstacles at the
maximum (high) for all obstacles.
2. We make the boundary to distinguish between barrier-related and non-barrier-related
non-innovators stricter by considering the average value for R&D innovators instead of
the average for all firms.
3. We make the boundary to distinguish between barrier-related and non-barrier-related
non-innovators stricter by considering the average value for R&D innovators instead of
the average for all firms and do not consider the answers to the “No need for
innovation” questions.
The results in table 8 compared the results in table 5 show clearly that the first two
modifications did not change the qualitative nature of the results. In contrast by using
definition 3 the magnitude of the coefficients for R&D innovators, non-technological
innovators and barrier-related non-innovators change considerably, as do the coefficients for
many other variables. Nevertheless, the signs of the coefficients remain the same.
Table 8: Robustness analysis: Regression results using modifications of the definition of barrier-related non-innovators, CIS 4
change in defintion of
barrier-related noninnovators
VARIABLES
log firm size
gazelle (y/n)
stable (y/n)
exporting (y/n)
part of foreign group
part of domestic group
manuf
medium-low innovation
medium innovation
medium-high innovatio
high innovation
country group 1
Country group 2
Country group 3
Basicness
Cumulativeness
R&D innovator
non-technology innova
barrier-related non-inn
Constant
Observations (weighte
pseudo R2
ll
exclusion of all firms that rank all obstacles at maximum in the
group of barrier-related non-innovators
lack of
lack of
lack of
technical
market
innovation
financial
knowledge knowledge
partners
barriers
skill barriers
reference for calculating the boundary ar not all firms but only R&D
innovators
lack of
lack of
lack of
technical
market
innovation
financial
knowledge knowledge
partners
barriers
skill barriers
reference for calculating the boundary ar not all firms but only R&D
innovators and questions regarding prior innovation were not
considered
lack of
lack of
lack of
technical
market
innovation
financial
knowledge knowledge
partners
barriers
skill barriers
-0.0090***
(-15.398)
0.0162***
(8.090)
-0.0076***
(-6.467)
0.0047***
(3.762)
-0.0517***
(-23.602)
-0.0184***
(-11.487)
0.0529***
(29.932)
-0.0539***
(-17.150)
-0.0381***
(-15.386)
-0.0341***
(-8.598)
-0.0444***
(-11.048)
-0.0483***
(-20.872)
-0.0754***
(-28.138)
0.0481***
(21.853)
0.0211
(1.368)
-0.1214***
(-16.601)
0.1317***
(75.019)
0.1247***
(89.915)
0.2549***
(167.357)
0.5020***
(30.323)
-0.0061***
(-10.642)
0.0085***
(4.292)
-0.0144***
(-12.397)
0.0046***
(3.772)
-0.0593***
(-27.421)
-0.0196***
(-12.416)
0.0585***
(33.546)
-0.0439***
(-14.145)
-0.0122***
(-4.996)
-0.0065*
(-1.652)
-0.0031
(-0.778)
-0.0385***
(-16.877)
-0.0581***
(-21.993)
0.0227***
(10.456)
0.0131
(0.862)
-0.1161***
(-16.090)
0.1284***
(74.113)
0.0914***
(66.863)
0.2438***
(162.313)
0.4742***
(29.039)
-0.0073***
(-12.808)
0.0081***
(4.166)
-0.0166***
(-14.486)
0.0185***
(15.256)
-0.0537***
(-25.151)
-0.0089***
(-5.686)
0.0343***
(19.897)
-0.0199***
(-6.513)
-0.0026
(-1.089)
0.0167***
(4.319)
0.0181***
(4.606)
-0.0912***
(-40.439)
-0.0931***
(-35.669)
-0.0224***
(-10.442)
0.1065***
(7.100)
-0.1573***
(-22.071)
0.1286***
(75.169)
0.0529***
(39.179)
0.2537***
(170.965)
0.5918***
(36.697)
-0.0152***
(-25.494)
0.0098***
(4.774)
-0.0287***
(-23.857)
0.0127***
(9.990)
-0.0728***
(-32.528)
-0.0067***
(-4.102)
0.0732***
(40.565)
-0.0431***
(-13.439)
-0.0214***
(-8.474)
-0.0218***
(-5.374)
0.0169***
(4.114)
-0.1664***
(-70.360)
-0.0961***
(-35.119)
-0.0000
(-0.010)
0.0694***
(4.416)
-0.1430***
(-19.145)
0.2024***
(112.878)
0.1304***
(92.127)
0.3388***
(217.870)
0.6382***
(37.755)
-0.0096***
(-15.381)
0.0255***
(11.848)
-0.0150***
(-11.831)
-0.0154***
(-11.535)
-0.0658***
(-27.974)
-0.0233***
(-13.550)
0.0607***
(31.966)
-0.0218***
(-6.458)
0.0089***
(3.356)
0.0257***
(6.042)
0.0507***
(11.733)
0.0028
(1.108)
-0.0665***
(-23.109)
0.0096***
(4.066)
-0.1596***
(-9.650)
-0.1337***
(-17.019)
0.1984***
(105.214)
0.1565***
(105.148)
0.3239***
(198.090)
0.6193***
(34.843)
-0.0088***
(-15.172)
0.0169***
(8.441)
-0.0071***
(-6.005)
0.0052***
(4.192)
-0.0519***
(-23.714)
-0.0188***
(-11.787)
0.0552***
(31.287)
-0.0574***
(-18.288)
-0.0404***
(-16.341)
-0.0397***
(-10.026)
-0.0491***
(-12.238)
-0.0477***
(-20.654)
-0.0746***
(-27.860)
0.0486***
(22.123)
0.0341**
(2.218)
-0.1183***
(-16.204)
0.1294***
(73.976)
0.1226***
(88.876)
0.2802***
(174.893)
0.4932***
(29.837)
-0.0060***
(-10.394)
0.0091***
(4.621)
-0.0139***
(-11.950)
0.0051***
(4.185)
-0.0594***
(-27.518)
-0.0200***
(-12.704)
0.0607***
(34.854)
-0.0472***
(-15.239)
-0.0144***
(-5.892)
-0.0118***
(-3.022)
-0.0076*
(-1.910)
-0.0380***
(-16.660)
-0.0573***
(-21.723)
0.0232***
(10.702)
0.0255*
(1.682)
-0.1131***
(-15.706)
0.1264***
(73.268)
0.0897***
(65.949)
0.2695***
(170.625)
0.4653***
(28.551)
-0.0071***
(-12.507)
0.0088***
(4.498)
-0.0160***
(-14.003)
0.0190***
(15.719)
-0.0537***
(-25.211)
-0.0093***
(-5.961)
0.0364***
(21.218)
-0.0233***
(-7.640)
-0.0048**
(-1.995)
0.0112***
(2.897)
0.0134***
(3.436)
-0.0906***
(-40.281)
-0.0923***
(-35.452)
-0.0219***
(-10.229)
0.1193***
(7.973)
-0.1541***
(-21.683)
0.1271***
(74.676)
0.0517***
(38.524)
0.2840***
(182.139)
0.5818***
(36.170)
-0.0150***
(-25.277)
0.0107***
(5.242)
-0.0280***
(-23.322)
0.0134***
(10.571)
-0.0731***
(-32.749)
-0.0073***
(-4.493)
0.0764***
(42.396)
-0.0478***
(-14.928)
-0.0246***
(-9.733)
-0.0292***
(-7.233)
0.0106***
(2.590)
-0.1656***
(-70.185)
-0.0949***
(-34.753)
0.0006
(0.289)
0.0869***
(5.539)
-0.1390***
(-18.645)
0.1986***
(111.256)
0.1270***
(90.198)
0.3684***
(225.346)
0.6274***
(37.195)
-0.0095***
(-15.254)
0.0265***
(12.300)
-0.0144***
(-11.360)
-0.0147***
(-11.015)
-0.0664***
(-28.239)
-0.0239***
(-13.942)
0.0639***
(33.653)
-0.0264***
(-7.817)
0.0058**
(2.169)
0.0185***
(4.355)
0.0446***
(10.327)
0.0035
(1.399)
-0.0653***
(-22.683)
0.0102***
(4.333)
-0.1426***
(-8.627)
-0.1299***
(-16.551)
0.1933***
(102.799)
0.1518***
(102.350)
0.3437***
(199.553)
0.6109***
(34.383)
-0.0054***
(-10.033)
0.0145***
(7.807)
-0.0020*
(-1.848)
0.0019*
(1.648)
-0.0355***
(-17.500)
-0.0159***
(-10.741)
0.0463***
(28.279)
-0.0508***
(-17.439)
-0.0398***
(-17.351)
-0.0433***
(-11.807)
-0.0513***
(-13.770)
-0.0422***
(-19.671)
-0.0744***
(-29.967)
0.0361***
(17.708)
0.0687***
(4.820)
-0.1069***
(-15.780)
0.2498***
(150.535)
0.2472***
(185.879)
0.4493***
(388.237)
0.3297***
(21.500)
-0.0025***
(-4.682)
0.0066***
(3.645)
-0.0088***
(-8.171)
0.0018
(1.600)
-0.0428***
(-21.440)
-0.0170***
(-11.701)
0.0516***
(32.051)
-0.0404***
(-14.128)
-0.0136***
(-6.012)
-0.0154***
(-4.256)
-0.0096***
(-2.623)
-0.0324***
(-15.379)
-0.0573***
(-23.479)
0.0107***
(5.360)
0.0596***
(4.255)
-0.1015***
(-15.255)
0.2479***
(151.983)
0.2154***
(164.800)
0.4474***
(393.358)
0.3004***
(19.933)
-0.0037***
(-7.149)
0.0064***
(3.578)
-0.0111***
(-10.463)
0.0157***
(14.067)
-0.0377***
(-19.104)
-0.0064***
(-4.474)
0.0277***
(17.433)
-0.0168***
(-5.948)
-0.0043*
(-1.935)
0.0075**
(2.111)
0.0113***
(3.117)
-0.0851***
(-40.863)
-0.0921***
(-38.226)
-0.0343***
(-17.322)
0.1537***
(11.100)
-0.1428***
(-21.709)
0.2454***
(152.222)
0.1742***
(134.800)
0.4453***
(396.075)
0.4212***
(28.276)
-0.0115***
(-21.125)
0.0084***
(4.509)
-0.0226***
(-20.542)
0.0097***
(8.373)
-0.0560***
(-27.310)
-0.0043***
(-2.910)
0.0672***
(40.635)
-0.0409***
(-13.906)
-0.0248***
(-10.723)
-0.0337***
(-9.102)
0.0077**
(2.042)
-0.1593***
(-73.605)
-0.0942***
(-37.583)
-0.0138***
(-6.687)
0.1282***
(8.908)
-0.1267***
(-18.534)
0.3284***
(195.988)
0.2614***
(194.705)
0.5141***
(440.060)
0.4516***
(29.174)
-0.0057***
(-9.973)
0.0238***
(12.076)
-0.0087***
(-7.458)
-0.0185***
(-15.089)
-0.0480***
(-22.226)
-0.0207***
(-13.123)
0.0539***
(30.927)
-0.0189***
(-6.087)
0.0062**
(2.527)
0.0142***
(3.641)
0.0420***
(10.587)
0.0099***
(4.336)
-0.0650***
(-24.603)
-0.0043**
(-1.994)
-0.1021***
(-6.728)
-0.1168***
(-16.209)
0.3300***
(186.844)
0.2933***
(207.215)
0.5215***
(423.456)
0.4253***
(26.058)
707,397
0.066
-400980
707,392
0.055
-391220
707,400
0.060
-382417
707,373
0.118
-415681
707,434
0.074
-451300
707,397
0.069
-399742
707,392
0.059
-389888
707,400
0.065
-380528
707,373
0.122
-414132
707,434
0.074
-451025
707,397
0.199
-424998
707,392
0.196
-411331
707,400
0.199
-404431
707,373
0.261
-460098
707,434
0.220
-478310
Source: CIS 4 and 2006 data accessed at the safe centre. t-statistics in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Country group 1: Belgium, Denmark,
Germany, Finland, France, Iceland, Luxemburg, Norway, Sweden; Country group 2: Czech Republic, Estonia, Hungary, Slovenia, Slovak Republic, Ireland;
Country group 3: Spain, Italy, Portugal, Greece; Country group 4: Bulgaria, Lithuania, Latvia, Romania, Cyprus, Malta.
5 Discussion and Conclusions
This paper investigated the propensity to rank innovation barrier as relevant across a number
of EU member States. The findings confirm on the one hand the relevance of the distinction of
revealed and deterring barriers to innovation. The finding confirm the impression that
deterring barriers to innovation are not one-dimensional but that there seems to be a
systemic interrelationship between the different deterring barriers to innovation for barrierrelated non-innovators.
The impression emerges that innovation barriers across the board are lighter in countries
closer to the technological frontier, than for countries more distant from the frontier. With
regard to specific barriers to innovation we find that financing constraints to innovative
activity are assessed to be more relevant in countries distant to the frontier, while skill
constraints are more relevant in frontier countries.
With regard to firm characteristics the findings suggest that high-growth firms generally are
more hit by innovation barriers than firms that do not experience high growth.
(to be expanded)
– 26 –
References:
(to be completed)
* This paper is based on our research for the report “Barriers to internationalisation and growth of EU's innovative
companies” prepared for the European Commission, DG Enterprise in the Project Pro Inno INNO-Grips II in 2010. We
thank Bernd Ebersberger and Andreas Reinstaller for important comments and Cesar Santos from the European
Commission for his guidance during the project. We are particularly indebted to Sergiu Parvan from Eurostat, who
supported and very efficiently managed our stays at the Safe Centre at Eurostat in Luxembourg. The opinions
expressed and arguments used are the responsibility of the authors.
– 27 –
Appendix A: Industry classification based on appropriability, opportunity,
cumulativeness and entrepreneurship (Peneder 2010)
Peneder (2010) constructs an innovation classification based on Community Innovation Survey (CIS) micro data for
21 countries. He classifies firms on the basis of entrepreneurship types and technological regimes.
Entrepreneurship: The firm classification distinguishes between creative and adaptive entrepreneurship. Creative
entrepreneurs are characterised by firm specific innovations and can be further separated into firms producing: (i)
their own process innovations; (ii) their own, new-to-the-market product innovations; or (iii) both. All other firms are
characterised as adaptive entrepreneurs. Among these Peneder distinguishes a fourth group of technology
adopters, which create product innovations that are new to the firm, but not to the market, or produce process
innovations mainly in cooperation with other enterprises or institutions. Finally, he identifies a fifth, residual group of
adaptive entrepreneurs that pursue opportunities other than technological innovation.
'Technological regimes' are characterised in terms of opportunity, appropriability and cumulativeness conditions,
whose combination defines the particular knowledge and learning environments within which the firm operates.
Opportunity conditions: The classification distinguishes four firm conditions according to the perceived technological
opportunities demonstrated by the firm's innovation activity: (i) no opportunities - the firm neither performs intramural
R&D nor purchases external innovations; (ii) acquisition - the firm innovates only by purchasing external R&D,
machinery, or rights (patents, trademarks, etc.); (iii) intramural R&D - the firm undertakes its own R&D, but the ratio of
innovation expenditure to total turnover is less than 5%; and (iv) high R&D - the firm performs intramural R&D and its
share of innovation expenditures in total turnover is more than 5%.
Appropriability conditions: (i) strategic - for firms relying exclusively on secrecy, complexity of design, or lead-time
advantages to protect their innovations; (ii) formal (other than patents) - firms that use the registration of design
patterns, trademarks, or copyright; (iii) patenting (either as well as or without strategic or other formal methods of
protection); (iv) full arsenal - firms make use of all of the above three means of protection; (v) none - firms employ
none of these tools.
Degree of knowledge cumulativeness: CIS data do not provide direct measures of cumulativeness. Peneder (2010)
combines two aspects of the CIS data. First he differentiates according to the relative importance of internal vs.
external sources of information. Second, he applies contrasting identification rules depending on whether the firm
seems to be a technology leader or a technology follower. Thus, firms within the 'creative response' classifications of
entrepreneurship are characterised as operating within highly cumulative regimes if internal sources of knowledge
are more or at least as important as external sources, and as operating in low cumulative regimes if the firm draws
more on external than internal knowledge for its innovations. These identification rules are reversed for 'adaptive
entrepreneurship' type firms.
Based on these criteria Peneder (2010) identifies five industry groups according to their innovation intensity and the
underlying technological regime:
High innovation intensity: NACE 29, NACE 30, NACE 31, NACE 32, NACE 33, NACE 72, NACE 73
Medium-high innovation intensity: NACE 17, NACE 23, NACE 24, NACE 25, NACE 26, NACE 27, NACE 34, NACE 35,
NACE 64
Medium innovation intensity: NACE 20, NACE 21, NACE 28, NACE 36, NACE 62, NACE 65, NACE 74
Medium-low innovation intensity: NACE 10, NACE 11, NACE 15*, NACE 16, NACE 22, NACE 40, NACE 41, NACE 66
Low innovation intensity: NACE 14, NACE 18, NACE 19, NACE 37, NACE 51, NACE 60, NACE 61, NACE 63, NACE 67.
Appendix B: Country group regressions
VARIABLES
log firm size
gazelle (y/n)
stable (y/n)
exporting (y/n
part of foreign
part of domes
manuf
medium-low in
medium innov
medium-high
high innovatio
Basicness
Cumulativene
R&D innovato
non-technolog
barrier-related
Constant
Observations
pseudo R2
ll
country group 1
lack of
innovation
financial
partners
barriers
lack of
technical
knowledge
lack of
market
knowledge
-0.0164***
(-17.407)
-0.0005
(-0.140)
-0.0164***
(-8.796)
-0.0027
(-1.342)
-0.0161***
(-5.757)
-0.0082***
(-3.955)
0.0393***
(12.416)
-0.0735***
(-12.958)
-0.0281***
(-6.687)
-0.0143*
(-1.853)
-0.0160**
(-2.169)
0.0578**
(2.101)
-0.0750***
(-6.476)
0.1347***
(49.713)
0.1441***
(58.433)
0.2350***
(93.743)
0.3556***
(13.631)
-0.0140***
(-14.828)
0.0013
(0.370)
-0.0026
(-1.392)
-0.0045**
(-2.216)
-0.0260***
(-9.298)
-0.0010
(-0.460)
0.0446***
(14.117)
-0.0522***
(-9.201)
-0.0062
(-1.489)
0.0173**
(2.243)
0.0194***
(2.626)
-0.0190
(-0.691)
0.0072
(0.624)
0.1640***
(60.513)
0.1267***
(51.370)
0.2695***
(107.522)
0.1513***
(5.800)
-0.0119***
(-12.951)
0.0074**
(2.189)
0.0068***
(3.739)
0.0067***
(3.337)
-0.0410***
(-14.936)
-0.0091***
(-4.481)
0.0267***
(8.628)
-0.0426***
(-7.679)
-0.0275***
(-6.692)
-0.0091
(-1.206)
-0.0036
(-0.492)
0.2567***
(9.528)
-0.1006***
(-8.873)
0.1498***
(56.481)
0.1041***
(43.134)
0.2523***
(102.830)
0.3411***
(13.362)
-0.0174***
(-18.930)
0.0027
(0.808)
-0.0138***
(-7.580)
0.0097***
(4.860)
-0.0088***
(-3.216)
0.0288***
(14.185)
0.0387***
(12.511)
-0.0451***
(-8.137)
0.0038
(0.924)
0.0316***
(4.172)
0.0359***
(4.965)
-0.0766***
(-2.844)
-0.0431***
(-3.803)
0.1724***
(65.031)
0.0912***
(37.816)
0.2614***
(106.602)
0.2763***
(10.829)
221,374
0.053
-104525
221,364
0.066
-104490
221,361
0.062
-99759
221,350
0.068
-99601
country group 2
lack of
innovation
financial
partners
barriers
skill barriers
lack of
technical
knowledge
lack of
market
knowledge
-0.0156***
(-14.362)
0.0266***
(6.682)
-0.0182***
(-8.481)
-0.0360***
(-15.323)
-0.0597***
(-18.517)
-0.0157***
(-6.562)
0.0329***
(9.033)
0.0206***
(3.150)
0.0621***
(12.872)
0.1278***
(14.374)
0.1377***
(16.194)
-0.4201***
(-13.266)
-0.1274***
(-9.567)
0.2391***
(76.706)
0.2207***
(77.785)
0.3834***
(132.986)
0.6494***
(21.648)
-0.0125***
(-8.579)
0.0458***
(6.876)
-0.0189***
(-5.863)
0.0123***
(3.541)
-0.0473***
(-8.448)
0.0413***
(7.101)
0.0563***
(11.385)
-0.0179**
(-2.062)
-0.0008
(-0.104)
0.0078
(0.653)
0.0103
(0.847)
-0.0046
(-0.100)
-0.1824***
(-8.562)
0.1585***
(29.872)
0.0619***
(14.009)
0.1922***
(46.374)
0.5780***
(11.939)
-0.0166***
(-11.140)
0.0461***
(6.773)
-0.0268***
(-8.133)
0.0349***
(9.793)
-0.0648***
(-11.336)
0.0169***
(2.839)
0.0563***
(11.149)
-0.0475***
(-5.345)
-0.0137*
(-1.808)
-0.0211*
(-1.723)
0.0042
(0.340)
0.0688
(1.467)
-0.1195***
(-5.491)
0.1863***
(34.366)
0.0799***
(17.683)
0.2150***
(50.789)
0.4350***
(8.797)
-0.0070***
(-4.649)
0.0217***
(3.155)
-0.0310***
(-9.305)
0.0319***
(8.868)
-0.0745***
(-12.884)
0.0084
(1.404)
0.0330***
(6.460)
-0.0667***
(-7.426)
-0.0207***
(-2.706)
-0.0253**
(-2.043)
-0.0139
(-1.106)
0.2531***
(5.332)
-0.2394***
(-10.872)
0.1129***
(20.588)
0.0632***
(13.830)
0.2281***
(53.244)
0.6759***
(13.509)
-0.0121***
(-7.440)
0.0104
(1.405)
-0.0335***
(-9.321)
0.0359***
(9.257)
-0.1268***
(-20.364)
0.0174***
(2.678)
0.0371***
(6.741)
0.0162*
(1.670)
0.0170**
(2.067)
0.0569***
(4.271)
0.0552***
(4.086)
-0.0008
(-0.015)
-0.2532***
(-10.679)
0.2274***
(38.508)
0.1361***
(27.661)
0.3750***
(81.283)
0.7891***
(14.643)
221,384
0.101
-135513
65,508
0.052
-29197
65,513
0.064
-30600
65,524
0.057
-31369
65,508
0.121
-36225
country group 3
lack of
innovation
financial
partners
barriers
skill barriers
lack of
technical
knowledge
lack of
market
knowledge
-0.0158***
(-9.507)
0.0284***
(3.753)
-0.0248***
(-6.767)
0.0116***
(2.926)
-0.0590***
(-9.285)
0.0232***
(3.509)
0.0566***
(10.072)
-0.0321***
(-3.244)
0.0325***
(3.856)
0.0452***
(3.330)
0.0792***
(5.752)
-0.1871***
(-3.586)
-0.1151***
(-4.755)
0.2738***
(45.417)
0.1285***
(25.589)
0.2853***
(60.589)
0.5206***
(9.467)
-0.0023**
(-2.521)
0.0181***
(6.335)
0.0032*
(1.813)
0.0046**
(2.540)
-0.1171***
(-27.757)
-0.0316***
(-11.744)
0.0548***
(21.337)
-0.0394***
(-8.660)
-0.0483***
(-13.449)
-0.0350***
(-6.561)
-0.0601***
(-10.912)
-0.0164
(-0.767)
-0.1764***
(-15.963)
0.1218***
(46.025)
0.1260***
(65.919)
0.2912***
(132.062)
0.6663***
(26.858)
0.0027***
(3.055)
0.0075***
(2.689)
-0.0153***
(-8.965)
0.0017
(0.936)
-0.1218***
(-29.611)
-0.0453***
(-17.277)
0.0572***
(22.840)
-0.0258***
(-5.803)
-0.0110***
(-3.146)
0.0048
(0.928)
-0.0019
(-0.355)
0.0214
(1.026)
-0.2368***
(-21.963)
0.0851***
(32.960)
0.0743***
(39.858)
0.2466***
(114.631)
0.7645***
(31.590)
-0.0042***
(-4.780)
0.0068**
(2.493)
-0.0292***
(-17.339)
0.0216***
(12.372)
-0.0739***
(-18.257)
-0.0043*
(-1.665)
0.0353***
(14.326)
-0.0052
(-1.189)
0.0114***
(3.311)
0.0378***
(7.372)
0.0322***
(6.101)
0.0391*
(1.904)
-0.2089***
(-19.697)
0.1101***
(43.362)
0.0275***
(15.000)
0.2707***
(127.944)
0.6984***
(29.337)
-0.0157***
(-16.677)
0.0125***
(4.281)
-0.0331***
(-18.432)
0.0088***
(4.721)
-0.1328***
(-30.778)
-0.0401***
(-14.574)
0.0899***
(34.202)
-0.0515***
(-11.059)
-0.0417***
(-11.339)
-0.0552***
(-10.100)
0.0050
(0.885)
0.1651***
(7.543)
-0.1713***
(-15.152)
0.2166***
(80.044)
0.1454***
(74.339)
0.3739***
(165.808)
0.6822***
(26.883)
65,535
0.086
-37582
378,500
0.056
-236535
378,500
0.045
-227187
378,500
0.054
-220948
378,500
0.094
-245111
country group 4
lack of
innovation
financial
partners
barriers
skill barriers
lack of
technical
knowledge
lack of
market
knowledge
-0.0033***
(-3.527)
0.0213***
(7.222)
-0.0113***
(-6.215)
-0.0164***
(-8.748)
-0.0765***
(-17.589)
-0.0381***
(-13.725)
0.0752***
(28.380)
-0.0236***
(-5.031)
-0.0146***
(-3.940)
-0.0082
(-1.481)
0.0245***
(4.309)
-0.0692***
(-3.131)
-0.1784***
(-15.643)
0.1576***
(57.724)
0.1304***
(66.124)
0.3089***
(135.775)
0.7107***
(27.762)
-0.0044*
(-1.930)
0.0378***
(5.498)
0.0133**
(1.969)
0.0057
(0.931)
-0.0421***
(-3.623)
0.0055
(0.509)
0.0801***
(10.917)
-0.0401***
(-2.956)
-0.0211*
(-1.807)
-0.0270
(-1.411)
-0.0348*
(-1.658)
-0.1019
(-1.274)
0.3190***
(10.857)
0.1148***
(6.140)
0.1081***
(15.340)
0.2319***
(35.414)
-0.5513***
(-8.365)
-0.0037*
(-1.664)
0.0198***
(2.908)
0.0018
(0.262)
0.0287***
(4.691)
-0.0358***
(-3.106)
-0.0127
(-1.193)
0.0815***
(11.193)
-0.0177
(-1.318)
0.0136
(1.169)
-0.0018
(-0.097)
0.0078
(0.375)
-0.2003**
(-2.525)
0.2878***
(9.876)
0.1401***
(7.554)
0.1118***
(16.009)
0.2179***
(33.554)
-0.4751***
(-7.268)
-0.0106***
(-4.519)
0.0051
(0.723)
0.0038
(0.547)
0.0252***
(3.945)
-0.0845***
(-7.020)
-0.0502***
(-4.524)
0.0522***
(6.876)
0.0149
(1.062)
0.0314***
(2.594)
0.0424**
(2.140)
0.0518**
(2.379)
-0.2140***
(-2.583)
0.1617***
(5.317)
0.0934***
(4.823)
0.0961***
(13.173)
0.2623***
(38.698)
-0.1128*
(-1.654)
0.0006
(0.248)
-0.0049
(-0.671)
-0.0164**
(-2.284)
0.0355***
(5.411)
-0.1556***
(-12.593)
-0.0606***
(-5.324)
0.0548***
(7.029)
0.0564***
(3.905)
0.0485***
(3.900)
0.0587***
(2.887)
0.1037***
(4.643)
-0.2311***
(-2.717)
0.0071
(0.228)
0.2438***
(12.260)
0.1633***
(21.810)
0.4132***
(59.360)
0.2560***
(3.654)
0.0000
(0.011)
0.0466***
(6.291)
0.0006
(0.076)
-0.0028
(-0.427)
-0.0714***
(-5.699)
0.0364***
(3.156)
0.0448***
(5.668)
0.0317**
(2.165)
0.0550***
(4.361)
0.0578***
(2.804)
0.0979***
(4.320)
-0.6207***
(-7.197)
0.3392***
(10.710)
0.2834***
(14.056)
0.1723***
(22.698)
0.2909***
(41.221)
-0.4510***
(-6.348)
378,500
0.061
-248458
30,416
0.051
-17015
30,416
0.050
-16761
30,416
0.057
-18063
30,416
0.127
-18875
30,416
0.070
-19295
skill barriers
Source: CIS 4 data accessed at the safe centre. t-statistics in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Country group 1: Belgium, Denmark, Germany, Finland,
France, Iceland, Luxemburg, Norway, Sweden; Country group 2: Czech Republic, Estonia, Hungary, Slovenia, Slovak Republic, Ireland; Country group 3: Spain,
Italy, Portugal, Greece; Country group 4: Bulgaria, Lithuania, Latvia, Romania, Cyprus, Malta.
– 29 –
VARIABLES
log firm size
gazelle (y/n)
stable (y/n)
exporting (y/n
part of foreign
part of domes
manuf
medium-low in
medium innov
medium-high
high innovatio
Basicness
Cumulativene
R&D innovato
non-technolog
barrier-related
Constant
Observations
pseudo R2
ll
country group 1
lack of
innovation
financial
partners
barriers
lack of
technical
knowledge
lack of
market
knowledge
-0.0096***
(-3.849)
0.0004
(0.040)
-0.0101*
(-1.776)
-0.0162***
(-2.973)
0.0135**
(2.450)
-0.0477***
(-4.575)
0.0666***
(8.688)
-0.0050
(-0.369)
0.0134
(1.264)
0.0650***
(3.630)
0.0276*
(1.702)
-0.1892**
(-2.403)
-0.2524***
(-6.322)
0.1955***
(29.878)
0.0959***
(13.777)
0.1771***
(19.798)
0.7588***
(8.683)
-0.0091***
(-3.631)
-0.0139
(-1.344)
-0.0240***
(-4.181)
-0.0048
(-0.877)
0.0087
(1.562)
-0.0463***
(-4.420)
0.0215***
(2.799)
-0.0052
(-0.386)
0.0291***
(2.734)
0.0489***
(2.723)
0.0783***
(4.817)
-0.0246
(-0.312)
-0.2237***
(-5.581)
0.2211***
(33.655)
0.0963***
(13.773)
0.1591***
(17.738)
0.6543***
(7.459)
-0.0135***
(-5.447)
0.0127
(1.239)
-0.0273***
(-4.811)
0.0152***
(2.797)
-0.0267***
(-4.881)
-0.0368***
(-3.554)
0.0334***
(4.393)
-0.0183
(-1.366)
0.0216**
(2.053)
-0.0016
(-0.090)
0.0235
(1.464)
0.0452
(0.578)
-0.2372***
(-5.989)
0.1638***
(25.236)
0.0721***
(10.433)
0.2003***
(22.594)
0.7104***
(8.193)
-0.0174***
(-7.029)
-0.0379***
(-3.712)
-0.0481***
(-8.486)
0.0118**
(2.184)
-0.0226***
(-4.120)
-0.0180*
(-1.741)
0.0055
(0.730)
0.0285**
(2.125)
0.0461***
(4.384)
0.0574***
(3.231)
0.0938***
(5.846)
0.1502*
(1.925)
-0.3562***
(-9.001)
0.1752***
(27.013)
0.0690***
(9.998)
0.2443***
(27.567)
0.9703***
(11.200)
24,030
0.059
-10995
24,035
0.066
-11090
24,040
0.051
-10812
24,035
0.072
-10790
country group 2
lack of
innovation
financial
partners
barriers
skill barriers
lack of
technical
knowledge
lack of
market
knowledge
0.0184***
(6.296)
0.0200*
(1.656)
-0.0400***
(-5.978)
-0.0321***
(-5.025)
0.0055
(0.851)
-0.0794***
(-6.514)
0.0351***
(3.916)
0.0129
(0.816)
0.0356***
(2.867)
0.0898***
(4.285)
0.1223***
(6.453)
-0.4155***
(-4.509)
0.0173
(0.371)
0.3376***
(44.076)
0.2246***
(27.565)
0.3754***
(35.879)
0.1665
(1.625)
0.0065***
(4.106)
0.0036
(0.683)
0.0014
(0.367)
0.0146***
(3.883)
-0.0580***
(-10.277)
-0.0025
(-0.406)
0.0327***
(6.243)
-0.0235***
(-2.747)
0.0091
(1.264)
0.0158
(1.374)
-0.0085
(-0.793)
0.2023***
(4.150)
-0.2565***
(-9.007)
0.0903***
(16.477)
0.0642***
(13.872)
0.1911***
(41.942)
0.6534***
(10.188)
0.0011
(0.657)
-0.0162***
(-3.076)
0.0056
(1.401)
0.0141***
(3.708)
-0.0602***
(-10.571)
-0.0167***
(-2.683)
0.0405***
(7.666)
-0.0344***
(-3.983)
-0.0030
(-0.416)
-0.0064
(-0.547)
-0.0208*
(-1.914)
0.3233***
(6.569)
-0.2425***
(-8.436)
0.0990***
(17.894)
0.0634***
(13.577)
0.1786***
(38.832)
0.6205***
(9.585)
0.0022
(1.303)
-0.0003
(-0.048)
0.0092**
(2.231)
0.0178***
(4.505)
-0.0638***
(-10.785)
-0.0369***
(-5.712)
0.0294***
(5.356)
-0.0323***
(-3.606)
-0.0002
(-0.032)
0.0033
(0.269)
-0.0009
(-0.079)
0.3492***
(6.829)
-0.2781***
(-9.308)
0.0683***
(11.879)
0.0296***
(6.103)
0.2101***
(43.971)
0.7171***
(10.661)
-0.0064***
(-3.557)
-0.0073
(-1.222)
0.0151***
(3.363)
0.0375***
(8.716)
-0.1451***
(-22.533)
-0.0371***
(-5.279)
0.0441***
(7.381)
-0.0205**
(-2.104)
-0.0165**
(-2.003)
-0.0271**
(-2.064)
-0.0249**
(-2.023)
0.7363***
(13.229)
-0.5951***
(-18.305)
0.1693***
(27.069)
0.1239***
(23.463)
0.3836***
(73.763)
1.4438***
(19.724)
24,046
0.125
-14796
55,927
0.041
-23985
55,927
0.039
-24508
55,927
0.044
-26652
55,927
0.118
-31380
country group 3
lack of
innovation
financial
partners
barriers
skill barriers
lack of
technical
knowledge
lack of
market
knowledge
-0.0027
(-1.463)
0.0037
(0.599)
-0.0137***
(-2.961)
0.0210***
(4.723)
-0.0835***
(-12.556)
-0.0200***
(-2.753)
0.0500***
(8.114)
-0.0631***
(-6.268)
0.0158*
(1.859)
-0.0020
(-0.147)
0.0281**
(2.211)
0.2539***
(4.416)
-0.2692***
(-8.018)
0.2729***
(42.248)
0.1864***
(34.178)
0.3065***
(57.075)
0.7711***
(10.199)
-0.0088***
(-8.094)
0.0131***
(3.799)
-0.0045**
(-1.973)
0.0004
(0.158)
-0.1340***
(-23.963)
-0.0439***
(-12.642)
0.0932***
(27.158)
-0.0472***
(-8.472)
-0.0309***
(-7.249)
-0.0328***
(-5.134)
-0.0379***
(-5.737)
-0.0469*
(-1.653)
-0.1161***
(-8.363)
0.1579***
(37.118)
0.1508***
(58.836)
0.3512***
(123.198)
0.5474***
(17.680)
-0.0100***
(-9.462)
-0.0037
(-1.109)
-0.0084***
(-3.754)
0.0090***
(3.708)
-0.1198***
(-22.036)
-0.0485***
(-14.355)
0.0834***
(24.993)
-0.0576***
(-10.631)
-0.0235***
(-5.681)
-0.0376***
(-6.054)
-0.0181***
(-2.811)
0.1149***
(4.162)
-0.1943***
(-14.396)
0.1605***
(38.805)
0.1298***
(52.104)
0.3390***
(122.308)
0.6747***
(22.411)
-0.0020*
(-1.933)
0.0067**
(2.086)
-0.0187***
(-8.708)
0.0241***
(10.226)
-0.1091***
(-20.803)
-0.0170***
(-5.231)
0.0653***
(20.272)
-0.0133**
(-2.542)
0.0018
(0.459)
0.0203***
(3.393)
0.0235***
(3.791)
0.0364
(1.365)
-0.1583***
(-12.161)
0.1797***
(45.030)
0.0881***
(36.660)
0.2851***
(106.642)
0.5431***
(18.704)
-0.0097***
(-9.070)
0.0354***
(10.478)
-0.0088***
(-3.912)
0.0295***
(11.962)
-0.1556***
(-28.341)
-0.0507***
(-14.871)
0.1116***
(33.099)
-0.0170***
(-3.104)
-0.0332***
(-7.942)
-0.0232***
(-3.694)
0.0026
(0.403)
0.1538***
(5.516)
-0.0827***
(-6.073)
0.3169***
(75.881)
0.1877***
(74.595)
0.4681***
(167.248)
0.4104***
(13.501)
55,927
0.091
-33187
216,512
0.083
-135357
216,512
0.083
-129280
216,512
0.070
-121461
216,512
0.156
-131375
country group 4
lack of
innovation
financial
partners
barriers
skill barriers
lack of
technical
knowledge
lack of
market
knowledge
-0.0052***
(-4.690)
0.0163***
(4.682)
-0.0057**
(-2.466)
0.0037
(1.468)
-0.1534***
(-27.105)
-0.0470***
(-13.367)
0.1303***
(37.511)
-0.0216***
(-3.834)
-0.0228***
(-5.296)
-0.0222***
(-3.435)
0.0146**
(2.181)
-0.1062***
(-3.694)
-0.1024***
(-7.291)
0.1909***
(44.334)
0.1724***
(66.459)
0.3939***
(136.510)
0.5268***
(16.810)
-0.0130***
(-7.000)
0.0095
(1.620)
0.0183***
(3.258)
-0.0284***
(-5.827)
-0.0931***
(-9.095)
-0.0367***
(-3.824)
0.0568***
(9.840)
0.0111
(1.128)
0.0466***
(5.482)
0.0150
(1.159)
0.0333**
(2.408)
-0.2088***
(-3.294)
-0.1408***
(-3.614)
0.1435***
(14.390)
0.0859***
(14.351)
0.2634***
(52.307)
0.6466***
(7.443)
-0.0131***
(-7.120)
0.0087
(1.504)
0.0167***
(2.994)
-0.0127***
(-2.634)
-0.0652***
(-6.429)
-0.0027
(-0.286)
0.0712***
(12.450)
-0.0653***
(-6.680)
-0.0226***
(-2.678)
-0.0503***
(-3.915)
-0.0261*
(-1.903)
0.0733
(1.167)
-0.1562***
(-4.048)
0.1324***
(13.410)
0.0563***
(9.503)
0.2468***
(49.489)
0.6303***
(7.325)
-0.0131***
(-6.897)
-0.0078
(-1.299)
-0.0028
(-0.495)
-0.0162***
(-3.249)
-0.0598***
(-5.713)
0.0105
(1.068)
0.0482***
(8.153)
0.0168*
(1.662)
0.0484***
(5.569)
0.0779***
(5.875)
0.0822***
(5.819)
-0.2055***
(-3.171)
-0.1619***
(-4.064)
0.1342***
(13.167)
0.1081***
(17.672)
0.2648***
(51.428)
0.7164***
(8.064)
-0.0157***
(-8.464)
-0.0315***
(-5.382)
-0.0162***
(-2.884)
-0.0215***
(-4.419)
-0.1300***
(-12.717)
-0.0233**
(-2.431)
0.0599***
(10.387)
0.0073
(0.743)
0.0124
(1.458)
-0.0003
(-0.022)
0.0235*
(1.702)
0.2521***
(3.983)
-0.3930***
(-10.103)
0.2856***
(28.687)
0.2172***
(36.344)
0.4442***
(88.325)
1.2201***
(14.062)
-0.0190***
(-9.877)
0.0081
(1.338)
0.0524***
(8.975)
0.0109**
(2.161)
-0.1014***
(-9.546)
0.0112
(1.129)
0.0691***
(11.536)
0.0144
(1.406)
0.0490***
(5.554)
-0.0308**
(-2.294)
-0.0038
(-0.268)
-0.0776
(-1.179)
-0.0931**
(-2.303)
0.2200***
(21.262)
0.1335***
(21.500)
0.3076***
(58.878)
0.5861***
(6.501)
216,512
0.109
-137965
51,183
0.063
-32761
51,183
0.056
-32276
51,183
0.059
-33905
51,183
0.156
-32696
51,183
0.080
-34656
skill barriers
Source: CIS 2006 data accessed at the safe centre. t-statistics in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Country group 1: Belgium, Denmark, Germany,
Finland, France, Iceland, Luxemburg, Norway, Sweden; Country group 2: Czech Republic, Estonia, Hungary, Slovenia, Slovak Republic, Ireland; Country group
3: Spain, Italy, Portugal, Greece; Country group 4: Bulgaria, Lithuania, Latvia, Romania, Cyprus, Malta.